Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

Michael Klinge; Choimaa Dulamsuren; Florian Schneider; Stefan Erasmi; Uudus Bayarsaikhan; Daniela Sauer; Markus Hauck

doi:10.1186/s40663-021-00333-9

Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

Klinge, Michael; Dulamsuren, Choimaa; Schneider, Florian; Erasmi, Stefan; Bayarsaikhan, Uudus; Sauer, Daniela; Hauck, Markus 2021-08-10 00:00:00 Background: Forest distribution in the forest-steppe of Mongolia depends on relief, permafrost, and climate, and is highly sensitive to climate change and anthropogenic disturbance. Forest fires and logging decreased the forest area in the forest-steppe of Mongolia. The intention of this study was to identify the geoecological parameters that control forest distribution and living-tree biomass in this semi-arid environment. Based on these parameters, we aimed to delineate the area that forest might potentially occupy and to analyse the spatial patterns of actual and potential tree biomass. Methods: We used a combination of various geographic methods in conjunction with statistical analyses to identify the key parameters controlling forest distribution. In several field campaigns, we mapped tree biomass and ecological parameters in a study area within the Tarvagatai Nuruu National Park (central Mongolia). Forest areas, topographic parameters and vegetation indices were obtained from remote sensing data. Significant correlations between forest distribution and living-tree biomass on one hand, and topographic parameters, climate data, and environmental conditions on the other hand, were used to delineate the area of potential forest distribution and to estimate total living-tree biomass for this area. Results: Presence of forest on slopes was controlled by the factors elevation, aspect, slope, mean annual precipitation, and mean growing-season temperature. Combining these factors allowed for estimation of potential forest area but was less suitable for tree-biomass delineation. No significant differences in mean living-tree biomass existed between sites exposed to different local conditions with respect to forest fire, exploitation, and soil properties. Tree biomass was reduced at forest edges (defined as 30 m wide belt), in small fragmented and in large 9 2 forest stands. Tree biomass in the study area was 20 × 10 g (1,086 km forest area), whereas the potential tree 9 2 biomass would reach up to 65 × 10 g (> 3168 km ). * Correspondence: mklinge1@gwdg.de Department of Physical Geography, Institute of Geography, University of Göttingen, Goldschmidtstraße 5, 37077 Göttingen, Germany Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Klinge et al. Forest Ecosystems (2021) 8:55 Page 2 of 20 Conclusions: The obtained projection suggests that the potential forest area and tree biomass under the present climatic and geoecological conditions is three times that of the present forest area and biomass. Forest fires, which mostly affected large forest stands in the upper mountains, destroyed 43% of the forest area and 45% of the living- tree biomass in the study area over the period 1986–2017. Keywords: Biomass, Fire, Forest-steppe, Geoecological factors, Mongolia, Permafrost Background vicinity of grasslands contributed to forest degradation The Mongolian forest-steppe represents the transition and decrease of forest area (Tsogtbaatar 2004; Dulamsu- zone between the southern limit of the boreal forest in ren et al. 2014). Widespread logging activities after forest Central Asia and the dry region of the Gobi Desert. It is fires with the target to extract the remaining wood from characterized by semi-arid climate and is highly vulner- the forest are thought to delay reforestation due to able to climate change and land-use intensification mechanical damage to young trees (Sakamoto et al. (Poulter et al. 2013; Yang et al. 2016; Khansaritoreh 2021). et al. 2017b). Various ecological stress factors have re- Boreal forests represent an important organic carbon cently reduced Mongolia’s forest area and thus most pool and are thus important for the global climate likely tree biomass as well (Dulamsuren et al. 2010a, (Goodale et al. 2002; Pan et al. 2011). Although most of 2010b; Hansen et al. 2013). Mongolia’s boreal forests the organic carbon in the boreal zone is stored in soils harbor a unique biodiversity of forest species, which is (DeLuca and Boisvenue 2012; Shvidenko and Sche- regionally obliterated due to deforestation (Hauck et al. paschenko 2014; Mukhortova et al. 2015), a considerable 2014). Moreover, conversion of boreal forests to steppe amount of carbon is also stored in the living-tree bio- grassland is estimated to reduce the organic carbon mass. Typically, carbon stocks in the total tree biomass − 1 stock density by roughly 40% not only due to the re- of boreal forests amount to 40–80 Mg C∙ha (Jarvis moval of biomass, but also as the result of carbon losses et al. 2001; Luyssaert et al. 2007; Thurner et al. 2014). from the organic layer (Dulamsuren et al. 2016). Investigations on tree biomass in the Mongolian forest- Drought stress and resulting declines in wood produc- steppe have been carried out in the Altai Mountains, tion and forest regeneration were repeatedly reported, southern Khangai Mountains (Dulamsuren et al. 2016), especially for Siberian larch (L. sibirica Ledeb.), which northern Khangai Mountains (Dulamsuren et al. 2019), makes up approximately 80% of the total forest area and Khentei Mountains (Danilin 1995; Danilin and (Dulamsuren et al. 2009, 2010a, 2010c; Liu et al. 2013). Tsogt 2014). Obtained tree biomass values were in the − 1 Mongolia’s mean annual air temperature has increased range of 123–397 Mg∙ha and thus (due to Mongolia’s by 0.27 K per decade (in total 1.7 K) from 1940 to 2001 position in the southernmost boreal zone) in and beyond (Batima et al. 2005), which is clearly above the global the higher range typical for boreal forests Altogether, average of 0.12 K per decade from 1951 to 2012 (IPCC these studies point to a decrease of average tree biomass 2013). For the period 1940–2006, Dagvadorj et al. (2009) from the more humid north to the drier south of the reported a seasonal differentiated temperature increase Mongolian forest-steppe. At local scale, tree biomass in of 0.11 K per decade in summer and of 0.51 K per dec- the interior of L. sibirica forests exceeds that at the for- ade in winter. For the same period, the authors stated est edges (Dulamsuren et al. 2016). No consistent signifi- spatially varying trends of increasing and decreasing cant differences in tree biomass were found between precipitation. forest stands of varying sizes and between forests grow- In addition, devastating forest fires disturbed large for- ing in grassland- and forest-dominated areas of the est areas in Mongolia over the past decades (Goldammer forest-steppe (Dulamsuren et al. 2019). 2002; Hansen et al. 2013; Nyamjav et al. 2007). Local au- Logging, other kinds of forest use such as forest pas- thorities of the Tarvagatai Nuruu National Park in our ture, and fire-setting have reduced the forest area and study region stated that forest fires became more fre- tree biomass in Central Asia since prehistoric times quent since the 1990s, whereby the most severe fires oc- (Miehe et al. 2007, 2014; Unkelbach et al. 2017, 2019). curred in 1996 and 2002. Goldammer (2002) reported The impact of these activities can be evaluated by esti- that fire fighters with an air fleet were installed from mating the potential extent of forest area based on cli- 1969 until the 1990s, when financial support from Russia matic and topographic parameters (Klinge et al. 2015). ended. Thus, the extensive fire events in 1996 and 2002 The parameters precipitation, temperature and evapor- could not be fought as effectively as previous fires. Fur- ation control the spatial pattern of forest and steppe dis- thermore, a lack in systematic forest management, insuf- tribution in the semi-arid forest-steppe (Nyamjav et al. ficient control of logging and forest pasture in the 2007; Dulamsuren and Hauck 2008; Klinge et al. 2018). Klinge et al. Forest Ecosystems (2021) 8:55 Page 3 of 20 In addition, topographic position plays an important role, as The region has continental climate with cold semi-arid forests are generally limited to north-facing slopes (Klinge conditions (Fig. 2). The monthly mean temperatures at et al. 2015;Haisetal. 2016). Thus, relief is an important fac- Tosontsengel range between − 31.7 °C in January and tor for the existence, vigour and tree density of forests. In 14.7 °C in July. Most of the annual precipitation occurs addition to natural factors, the present forest distribution is during summer, from low-pressure cells brought by the strongly influenced by human impact that increased since westerlies (Batima et al. 2005). In contrast, the Siberian prehistoric times. Logging is done in an unsystematic man- High during winter causes mostly dry conditions. The ner for timber and fuelwood and pervasive in the forest- cold climate promotes discontinuous permafrost, with steppe. Its intensity has often increased after the transition permafrost mainly occurring in valley bottoms, upper from planned to market economy in the 1990s (Dulamsuren mountains, and partially on slopes. The existence of et al. 2014). Livestock kept by pastoral nomads influences permafrost ice requires some soil moisture, whereas dry forest regeneration at forest margin and in the interior of soil conditions lead to dry permafrost, i.e., perennially small forests. In the Mongolian Altai, increased livestock frozen ground without ice. densities promoted the establishment of L. sibirica seedlings The maximum altitudes of the study area of up to due to the creation of gaps in the ground vegetation, but later 3200 m a.s.l. occur in its southern part. They are charac- reduced the density of tree regeneration in the sapling stage, terised by mountain plateaus with cryoplanation terraces as the seedlings are a preferred diet of goats (Khishigjargal (Richter et al. 1963; Kowalkowski and Starkel 1984). et al. 2013). Goat numbers in Mongolia have multiplied since These highest regions above the upper treeline at the 1990s owing to the high economic significance of cash- approx. 2500 m a.s.l. belong to the periglacial belt, with mere wool for the herder households (Lkhagvadorj et al. alpine vegetation and bare, rock-debris covered land sur- 2013a, 2013b). Relative to livestock, browsing by wild ungu- faces (Klinge et al. 2018). In the northern part, the lates is of subordinate importance due to lower densities and mountains are lower, and mountain forest-steppe covers a large hunting pressure. The regeneration success of L. sibir- the north-facing slopes up to the summits. The main ica is primarily dependant on moisture availability and her- valleys run from south to north, leading into the east- bivory (Dulamsuren et al. 2008; Khishigjargal et al. 2013). It west running valley of the Ider Gol (Gol: Mongolian for is only loosely related to fire in the Mongolian forest steppe, River) at an elevation of 1600 m a.s.l.. The geological as the dry climate (supported by livestock) generates gaps in basement consists of Permian metamorphosed sedi- the ground vegetation where seedlings can establish (Danilin mentary and acid plutonic rock, and Carboniferous 1995;Dulamsurenet al. 2010b). Tree-ring chronologies often mafic rock (Academy of Sciences of Mongolia, Academy show annual tree establishment over longer periods with of Sciences of USSR, 1990). Coarse detritus of these bed- moist climate, but unrelated to fire (Dulamsuren et al. 2010a, rocks forms slope debris, which is often mixed with and 2010c; Khansaritoreh et al. 2017a). covered by sandy to silty aeolian deposits. Based on the state of knowledge described above, we Dense, extensive forests occur south of the Ider Gol, addressed the following hypotheses: whereas north of the river, forests are more fragmented and steppe vegetation is dominant (Dulamsuren et al. (I) Climatic and topographic parameters limiting the 2019). A clear spatial pattern of forests (made up of L. general distribution of larch forests in the study sibirica) on north-facing slopes and steppe on south- area can be deduced by spatial analysis of remote facing slopes is typical in the forest-steppe of Mongolia sensing data. (Hilbig 1995; Treter 1996). This vegetation pattern is (II) The combined effect of additional environmental generally controlled by low precipitation (< 300 mm), factors (differing from the topo-climatic factors high evapotranspiration and relief-controlled differences controlling forest distribution) control the actual in insolation in the mid-latitudes (Schlütz et al. 2008; living-tree biomass in the study area. Hais et al. 2016). Riverine forests consist of willow (III)Frequent forest fires, logging, and wood pasture (Salix), poplar (Populus), and larch (L. sibirica). Since strongly reduced the forest area and living-tree bio- these alluvial forests are supported by groundwater, they mass since prehistorical time. Thus, forests only are rather independent from local precipitation. Pleisto- partially cover the potential forest area that can be cene dune fields with scattered individual old larch trees deduced from climatic and topographic conditions. are abundant in the basins. Many local forest and steppe fires occur during summer (Goldammer 2007; Hessl Materials and methods et al. 2012). Severe forest fires in 1996 and 2002 Study area destroyed extensive forests. Many of these former forest The study area is located on the northern edge of the areas have not yet regrown. A timber factory and forest Khangai Mountains near the town Tosontsengel in tracks were established in the Tosontsengel region dur- northern central Mongolia (98°16′ E, 48°46′ N) (Fig. 1). ing Soviet times to facilitate intensified forest Klinge et al. Forest Ecosystems (2021) 8:55 Page 4 of 20 Fig. 1 Study area. a) Overview of Mongolia with position of the map shown in b) (black rectangle). b) Location of the study area in the forest- steppe of northern central Mongolia. Forest distribution was adapted from Klinge et al. (2018), burnt forest area (2000–2018) was adapted from Hansen et al. (2013). The digital elevation model (DEM) was created from SRTM (Shuttle Radar Topography Mission) data. The black rectangle in b) indicates the position of the image shown in c). c) True-colour satellite image of the study area near Tosontsengel (Landsat 8, September 22, 2014) Klinge et al. Forest Ecosystems (2021) 8:55 Page 5 of 20 Fig. 2 Climate of the study area around the town Tosontsengel (black circle). Data from the CHELSA V1.2 dataset, measuring period 1979–2013 (Karger et al. 2018), the shaded relief illustration is based on TanDEM-X data exploitation since the 1960s. Former clear-cutting is still 23, 1986 to delineate the distribution of forest prior to documented by rotted tree stumps inside the forests. In- extensive forest destruction through fires. This image dustrial logging was abandoned after the political change from 1986 was the best image for the period before the in the early 1990s, but has been resumed to some extent. onset of extensive forest fires. We determined the actual Illegal logging happens selectively inside of forests and forest area by integrating several scenes of Landsat 8 affects individual remnants in burnt areas, but no exten- (May 14, 2013; June 20, 2015) and Sentinel 2 (Sep 14, sive clear-cutting occurs. In addition, the local popula- 2016; Sep 19, 2017). After the intersection of the classifi- tion extracts selectively fuelwood from the forests. cation, we went through an intensive visual check of all forest polygons to delete and edit wrong classified areas. Remote sensing analysis of forest distribution, forest In the forest-steppe, distinct boundaries between forest categories and landscape units patches and grassland allowed for highly accurate forest This work step was crucial, as the main aim of this study maps. As shown by Klinge et al. (2015), image classifica- was to identify a possible mismatch between actual and tion combined with manual post-editing leads to an potential forest area. Input data for this work step in- overall accuracy of > 0.99. We used the difference in for- cluded Landsat 5, Landsat 8 and Sentinel 2 images est area between the images of 1986 and 2017 to work (Fig. 3, first line). For the mapping of forest areas, we ap- out the burnt forest area for this period. Because of the plied a semi-automatic approach after Klinge et al. different spatial resolutions of the satellite images, diver- (2015). The manual mapping of forest stands from satel- gences less than 20 m at the forest edges were neglected. lite imagery was supported by a previous supervised We distinguished several forest categories that occur maximum likelihood classification, where training sam- in different landscape units. Based on the proportion be- ples were distinguished for forest, steppe and water bod- tween forest and steppe, we differentiated between forest ies. We used a Landsat 5 satellite image from September stands in forest-dominated area and steppe-dominated Klinge et al. Forest Ecosystems (2021) 8:55 Page 6 of 20 Fig. 3 Workflow of this study area. The area, where the mountain tops reach above the slope, and insolation. We applied a GIS tool to estimate upper treeline, was defined as high-mountain area. For- the cumulative solar radiation input for the period Mai- ests in flat areas (< 2°) along rivers were interpreted as September 2017, which served as mean growing season alluvial forests. Forest stands on dunes were directly (MGS). These parameters were extracted for forest area identified in the satellite images. Furthermore, we distin- using the forest map of 1986, in order to identify rela- guished four forest-size classes in the forest-dominated tionships between potential forest distribution and relief. and steppe-dominated areas, respectively: F1, G1 ≤ 0.1 This approach allowed us to determine suitable topo- 2 2 2 2 km ; F2 = 0.1–1km ;F3=1–5km ;F4 ≥ 5km , using a graphic conditions and topographic thresholds for forest spatial buffer of 30 m to distinguish the forest edges growth from the respective value range. We produced a from the interior. These classification schemes were map of potential forest area based on relief parameters adapted from Dulamsuren et al. (2016, 2019). (PFA ), assuming that forest growth is possible in all areas within the topographic thresholds for forest Determination of suitable topographic conditions and growth, applying the approach of Klinge et al. (2015). topographic thresholds for forest growth Relief leads to variations of local climate, which is a par- Determination of suitable climatic conditions and climatic ticularly important aspect in semiarid central Asia thresholds for forest growth (Klinge et al. 2015). Therefore, we created a digital eleva- Similarly to the determination of PFA described above, tion model (DEM) based on TanDEM-X data (Fig. 3, we also analysed PFA based on relationships between first line). Due to high horizontal (10 m × 10 m) and ver- forest distribution and climatic parameters (PFA ). For tical (< 1 m) resolution of the DEM, the calculated ter- this purpose, we used climate data from the CHELSA rain surface was distorted by the forest canopy, V1.2 dataset (Karger et al. 2018), which we resampled especially at the edges of the forest stands. Therefore, we from the originally 30-arcsec resolution to obtain 30-m used the map of the actual forest area obtained from sat- resolution (Fig. 3, first box). This reanalysed climate ellite imagery to correct the DEM in forests. dataset enabled us to consider terrain parameters and The corrected DEM allowed us to extract various wind effect, and thus allowed us to obtain an improved topographic parameters, including elevation, aspect, representation of climate conditions in relief terrain Klinge et al. Forest Ecosystems (2021) 8:55 Page 7 of 20 (Karger et al. 2017). We calculated mean annual precipi- aboveground and belowground living-tree biomass. Dif- tation (MAP), mean annual potential evaporation, and ferences in the estimates of the two functions are dis- mean growing-season temperature (MGST, as the aver- cussed in Dulamsuren et al. (2016). We presumed that age of the monthly mean temperatures between May the increase of total tree biomass over the four-year and September) for the period 1973–2013. Temperature period of biomass-data collection in the field was less and precipitation were largely independent (Fig. 2), than the precision of the allometric method. Forest whereby temperature followed a vertical gradient, and stands where tree stumps indicated logging, were precipitation showed an additional longitudinal gradient mapped as forests with “logging”, whereas forest stands caused by the westerlies. Potential evaporation was so without tree stumps were mapped as forests with “no closely correlated with MGST that we did not include it logging”. Forest stands, where burnt bark and/or charred as an additional parameter for delineating PFA . wood indicated former fire events, were mapped as hav- The climatic parameters were extracted for forest area ing “fire indicators”, those without as having “no fire in- using the forest classification of 1986, in order to iden- dicators”. Ground vegetation structure, soil profiles and tify relationships between potential forest distribution detection of permafrost provided auxiliary data. Ground- and climate, and to determine suitable climatic condi- vegetation structure was important for assessing, which tions and climatic thresholds for forest growth. The portion of the NDVI of the forest sites was contributed intention was to produce a map of potential forest area by ground vegetation, because tree-canopy closure was based on climatic parameters (PFA ), assuming that for- less than 53%. In 24 of the plots, we measured leaf area est growth is possible in all areas within the climatic index (LAI) using a LI-COR Plant Canopy Analyzer thresholds for forest growth, thus, using the same ap- LAI-2200 C (Licor Biosciences, USA). Soil profiles were proach as for the PFA map. used to distinguish soils developed in sandy sediment and slope debris, and to detect permafrost, as permafrost Tree-biomass analysis is a crucial factor for forest distribution due to its During fieldwork in the years 2014–2018, we determined impounding effect for meltwater from seasonal ground living-tree biomass on 20 m × 20 m plots (Fig. 3, last box ice, keeping the meltwater available for trees. Permafrost in first line), by measuring tree diameter at breast height distribution was not used for biomass and PFA delinea- (dbh) and tree height of all living trees exceeding a tion, but its ecological feedback represents a relevant height of 4 m. In addition, we counted seedlings, saplings secondary parameter as shown by Klinge et al. (2021). and trees < 3 m, whereas we did not include the dead Classification of the plot data according to other influ- biomass. We used either a Vertex IV ultrasonic clinom- encing factors (Table 1) in addition to relief and climate, eter and T3 transponder (Haglöf, Långsele, Sweden) or a allowed for extracting effects of these factors on tree True Pulse 200 laser rangefinder (Laser Technology, biomass through statistical analysis. Inc., USA) for measuring tree height. Stem diameter was Dulamsuren et al. (2019) already analysed tree biomass calculated from stem circumference as measured with a in the interior of larch forests on slopes in the same measuring tape. Plot dimensions were determined by study area, thereby focusing on larch stands in the measuring tape, and plot-corner positions were mea- optimum stage of the forest development cycle (Jacob sured by GPS at an accuracy of ~ 3 m. For spatial refer- et al. 2013; Feldmann et al. 2018) and excluding distur- ence of the plot data, the centre between the four plot bances from fire or logging. Complementary to that corners was calculated. For statistical correlation analysis study, we also included larch stands influenced by vari- between biomass and remote sensing data, mean values ous factors, in order to also address the response of tree were interpolated of the pixels located within a 400-m biomass to these factors. In doing so, we also tested the circle around this point (Fig. 3, upper left side). In this potential of remote sensing techniques for upscaling way, we analysed 140 plots, including forest- and steppe- plot-based data to the landscape level. We used tree- dominated landscapes, forest edges and interiors, toe biomass data of 30 L. sibirica plots on slopes from slopes, mid-slopes, upper slopes, pristine and exploited Dulamsuren et al. (2019) and added tree-biomass data forests, as well as forest stands on slopes of different as- from forest edges and further L. sibirica plots of differ- pects and of different forest-stand sizes. We selected the ent forest-stand sizes (classes F1/G1 to F4) and forest- biomass plots according to their representativeness for a to-grassland ratios (forest-dominated area with classes larger surrounding area, to minimise discrepancies be- F1 to F4 vs. steppe-dominated area with class G1). tween field data and remote sensing data. In addition to these forests on slopes, we also analysed We applied the two allometric functions for Siberian larch forests on alluvial sand in floodplains. However, larch (L. sibirica) in Mongolia published by Battulga their limited size did not allow for obtaining separate et al. (2013) and Dulamsuren et al. (2016), and used the datasets for forest interior and edge. Altogether, we dis- mean of the results of both equations to estimate the tinguished 12 larch-stand categories, including the Klinge et al. Forest Ecosystems (2021) 8:55 Page 8 of 20 Table. 1 Mean living-tree biomass (above and belowground) for different forest categories affected by various factors. Plots, where a possible influence of a certain factor could not be assured, were excluded from the respective part of the analysis. SE = standard error, n = number of plots F1 SE n F2 SE n F3 SE n F4 SE n G1 SE n Total 198.7 11.2 29 208.0 11.8 31 212.5 13.9 31 182.0 12.9 34 142.2 10.7 10 Forest interior 219.3 14.0 18 218.2 12.6 26 220.6 13.4 26 181.7 13.8 30 145.4 15.0 7 Forest edge 165.1 13.7 11 155.2 21.0 5 170.7 46.2 5 184.3 35.2 4 134.8 2.7 3 Difference 54.2 63.0 49.9 −2.6 10.6 No fire indicators; 211.5 22.3 2 210.7 18.1 5 219.0 12.8 4 170.7 14.3 15 no logging No fire indicators 204.8 18.7 14 206.2 12.8 14 194.3 17.4 21 168.3 13.3 23 122.1 2.7 2 Fire indicators 179.5 15.2 9 171.6 20.9 11 231.3 14.4 6 175.5 39.7 5 154.9 1 Difference 25.2 34.6 −36.9 −7.2 No logging 211.5 22.3 2 234.3 11.9 8 238.8 23.8 6 180.4 13.7 18 134.1 4.0 2 Logging 197.8 11.9 27 198.9 14.4 23 207.4 16.6 24 183.8 22.7 16 144.3 13.2 8 Difference 13.7 35.4 31.4 −3.4 Slope debris 211.0 18.4 13 207.8 16.5 18 196.3 18.5 19 176.7 14.3 28 Sand layer 188.8 13.3 16 204.5 15.8 14 238.3 18.3 12 214.4 24.5 7 Difference 22.2 3.4 −42.0 −37.7 additional influencing factors forest interior / forest remnants. In the forest-dominated central and north- edges of the slope-forest categories F1, F2, F3, F4 and eastern parts of the study area, the most extensive burnt G1, and the floodplain forests as independent variables, forest areas were in the upper mountains. Only few for- and differentiating between logged / not logged and est stands burnt down in the steppe-dominated areas in burnt / not burnt forest stands as covariates, based on the north-western and eastern parts of the study area. the presence / absence of tree stumps and fire scars. Usual windthrow that creates single deadwood inside Permafrost distribution natural forests was not considered as disturbance. Permafrost was restricted to large forest stands on slopes We calculated the mean living-tree biomass for each in the forest-dominated area and high-mountain area, as forest category (as affected by the diverse factors), con- observed in our soil profiles (Klinge et al. 2021). Under sidering, e.g., logging, fire indicators, and topsoil condi- large forest stands (forest-size class F4) on north-facing tions (Table 1, Table S1). We checked the tree-biomass slopes, the permafrost was rich in ice and occurred data of each forest category for normal distribution and already at shallow depth, whereas east- and west-facing tested the differences in living-tree biomass between the slopes had only small patches of permafrost that started forest categories for statistical significance using Dun- at depths of more than 1 m. There was no field evidence can’s multiple range test calculated in SPSS. for permafrost under fragmented forest stands (G1, F1, We multiplied the area of each forest category with F2) and burnt forests (Klinge et al. 2021). the mean tree biomass of that forest category, using three scenarios, namely i) the actual forest area, ii) the PFA delineation based on relief parameters (PFA ) forest area of 1986, and iii) the potential forest area The upper treeline in the study area rises from 2400 m (PFA). Delineation of potential alluvial-forest area was in the north to 2600 m a.s.l. in the south (Klinge et al. not feasible, because the alluvial-forest distribution pat- 2018). Since 1986, forest fires in the upper mountains tern was largely controlled by the erosion-deposition dy- led to a decline in the mean elevation range of forests namics of the braided rivers. (95%) to 1600–2400 m a.s.l.. South-facing slopes in the forest-steppe are generally covered by steppe vegetation; Results they may be partially forested above 2100 m a.s.l. (Fig. 5). Spatial patterns of forest-fire and permafrost distribution Forests also occur in areas of maximum MGS insolation, Forest-fire distribution which demonstrates that insolation is no limiting factor The high-mountain area in the southern part of the for forest growth in the study area. study area lost the largest portion of forest through fire We delineated the PFA by clipping the area, where all over the past decades (Fig. 4). Its formerly large forests three parameters ‘aspect’ (no forest below 2100 m a.s.l. turned into numerous small and fragmented forest on slopes with aspect 135°–225°), ‘slope gradient’ (0– Klinge et al. Forest Ecosystems (2021) 8:55 Page 9 of 20 Fig. 4 Landscape units in the study area, with actual forest area and burnt forest area (reference year 2018). The shaded relief illustration is based on TanDEM-X data 25°), and ‘elevation’ (upper treeline: 2400 m a.s.l. in the presently forested area on steep slopes, from the valleys north, 2600 m a.s.l. in the south) allowed for tree growth up to the upper treeline at 2600 m a.s.l.. (Fig. 3). The most striking outcome of this PFA projec- tion was a much more extensive forest cover on toe PFA delineation based on climatic parameters (PFA ) slopes and pediments, which are at present generally The spatial resampling of the climate data by linear covered by steppe vegetation (Fig. 6). In the high- interpolation produced some noise, because small topo- mountain area, the estimated PFA exceeded the graphic variations could not be considered. Therefore, r Klinge et al. Forest Ecosystems (2021) 8:55 Page 10 of 20 Fig. 5 Top - Change of maximum MGS insolation (left axis) with elevation (MGS = mean growing season, May–September). Solid lines = MGS insolation on forest area, dashed lines = MGS insolation on the total land surface. Bottom - Forest area in hectares (right axis) in 1986 plotted against elevation. * The forest area on south-exposed slopes is shown in ha × 10 we did not use the obtained climate dataset to deduce except for a difference of the G1 forest edge plots to the the climatic thresholds for forest growth by histogram F1 and F2 interior plots (Fig. 8). Forests of the size clas- − 1 analysis. Instead, we derived these thresholds from step- ses F1, F2, and F3 had 50–63 Mg·ha less living-tree wise adjustment of the climatic thresholds until the biomass at forest edges than in their interiors; only the spatial pattern of temperature and precipitation with forest-size class F4 showed no distinct difference in tree positive conditions for tree growth included all forest biomass between forest edges and interior (Table 1). stands of 1986. The climatic thresholds obtained from Small fragmented forests G1 in the steppe-dominated − 1 this approach are listed in Table 2. The obtained PFA area had up to 70 Mg·ha less tree biomass than those showed larger forest areas on the upper south-facing in the forest-dominated area. Logging, forest fire and slopes and small flat summits. It matched well with the sediment type did not significantly influence living-tree upper treeline in the high-mountain area (Fig. 7). Com- biomass. The large proportion of forest with logging in pared to the PFA , the PFA did not extend as far down the categories of small fragmented forests (size classes r c into the basins, which may be due to low precipitation F1 and G1) pointed to a higher exploitation pressure on there. these small forests compared to larger stand sizes. Plot-based tree-biomass data Remote-sensing analysis based on tree-biomass estimates Mean tree height ranged between 12 and 20 m, whereas We tested several NDVI datasets and various topo- the maximum heights of single trees reached up to 32.7 graphic parameters for significant single or multi- 2 − 1 m. Stand basal area ranged from 5 to 91 m ·ha , with correlations with measured tree biomass, as such corre- 2 − 1 an average of 38.8 m ·ha . Mean tree ages were 100– lations would allow for interpolating tree biomass to 200 years, whereby maximum tree ages reached up to landscape scale. However, we found no statistically sig- 380–413 years. Living-tree biomass in larch forests on nificant correlations (r > 0.5, p < 0.05). One problem for − 1 slopes ranged between 25 and 380 Mg·ha . Maximum NDVI analysis was the low number of multispectral sat- − 1 tree biomasses of 440–688 Mg·ha were found in for- ellite images with < 10% cloud cover taken during sum- ests on floodplains, whereas larch trees on sand dunes mer. Another problem was the weak correlation only formed open woodlands with less tree biomass (48 between needle volume and tree biomass, as leaves and − 1 Mg·ha , n = 1). In all stand-size classes of the forests of needles provide the chlorophyll signal in multispectral the forest-dominated area, tree-biomass means and me- satellite images. Danilin and Tsogt (2014) stated that − 1 dians were within the range 180–220 Mg·ha ; max- needle biomass is independent of the average age of a − 1 imum tree biomass exceeded 320 Mg·ha (Table 1). larch stand, whereas tree biomass increases with tree Duncan’s multiple range test did not proof statistically age. The absence of a significant correlation between leaf significant differences between the forest categories, area index (LAI) and tree biomass measured on 24 plots Klinge et al. Forest Ecosystems (2021) 8:55 Page 11 of 20 Fig. 6 Forest distribution in 1986 and potential forest areas (PFA) delineated based on climate and relief. The shaded relief illustration is based on TanDEM-X data confirmed this statement (Fig. 9). Overall, the statistical Table. 2 Thresholds of mean growing-season temperature relationships between NDVI, needle volume and tree (MGST) and mean annual precipitation (MAP) used for PFA c biomass were poor. delineation Aspect Synthesis of the results of the forest-distribution and North East South West tree-biomass assessments The size of the study area was 6355 km . A closed third Maximum MGST (°C) 10.8 10.4 8.6 10.0 of the total area (1898 km ) was still forested in 1986. Minimum MGST (°C) 6.5 6.5 6.5 6.5 Since then, the forested area declined to 1086 km (ac- Maximum MAP (mm) 320 310 340 340 tual forest area). The delineated PFA yielded 3168 Minimum MAP (mm) 160 170 290 165 (PFA ) and 3553 km (PFA ), respectively. Details on the c r Klinge et al. Forest Ecosystems (2021) 8:55 Page 12 of 20 Fig. 7 Frequency-distribution curves of mean growing-season temperature (MGST) and mean annual precipitation (MAP) in the study area. Solid lines = forest area, dashed lines = total area. Climate data source: CHELSA V1.2 (Karger et al. 2017), period 1979–2013, spatially resampled to 30 m by linear interpolation differences between actual forest area, forest area in forest-dominated area, because both PFA projections 1986, and PFA, are listed in Table 3. The largest portion had resulted in forest-dominated landscapes (Table 5). of the actual forest area falls into the forest-size class F4 The actual tree biomass in the study area was 57% of the (8.6%) in the forest-dominated area. Prior to the large one in 1986, corresponding to 30% and 34% of the tree fire events, this forest-size class was also widespread in biomass estimated for the PFA and PFA , respectively r c the high-mountain area (9.2%). Altogether, a forest area (Table 5). The greatest losses of living-tree biomass due of 812.5 km (12.8% of the total study area and 42.8% of to forest fires since 1986 were detected in the large for- the formerly forested area) was destroyed by fire since ests (size class F4) of the forest-dominated area and the end of the last century. The burnt forest area was high-mountain area, whereas tree-biomass losses negligible in the steppe-dominated area and small in the through fire were less severe in the steppe-dominated forest-dominated area, but it amounted up to 95% of the area and alluvial forests. formerly forested area in the high-mountain area. The ratios between forest interior and forest edge (F /F ) for Discussion i e each forest-size class were relatively consistent for the Forest distribution different landscape units (Table 3). The main natural factors that control the spatial distri- Due to the low correlation between topographic pa- bution and vigour of forests in the Mongolian forest- rameters, NDVI and living-tree biomass (Fig. 9), estima- steppe are low precipitation and high evapotranspiration. tion of the total living-tree biomass in the study area by The latter depends on insolation, which in turn varies use of regression functions was not feasible. Therefore, with relief, resulting in a lack of forests on south-facing we estimated the total living-tree biomass by multiplying slopes (Dulamsuren and Hauck 2008; Hais et al. 2016; the specific mean tree biomass of each forest category by Klinge et al. 2018). Where forests occur, the forest can- the area of that forest category in 1986 and 2018 (Ta- opy fosters dense ground vegetation and an organic sur- bles 3, 4, 5). The total living-tree biomass of the PFA face layer that insulates the soil from warm air during was calculated based on the mean living-tree biomass of summer (Dashtseren et al. 2014). In this way, forests Klinge et al. Forest Ecosystems (2021) 8:55 Page 13 of 20 − 1 Fig. 8 Boxplots of tree biomass (Mg·ha ) of forests differing with respect to edge effects, influence of fire, logging and sediment type. Horizontal line = median, bars = quartiles, whiskers = range, dots = outliers. Means sharing a common letter, do not differ significantly (Duncan’s multiple range test, p ≤ 0.05) support discontinuous permafrost (Klinge et al. 2021). In Nevertheless, steppe vegetation predominates on the turn, permafrost helps trees to survive summer droughts, pediments, mainly because of herbivore grazing (Hilbig as it prevents meltwater that is released above the 1995). The climate-limited potential forest area (PFA ) permafrost table from percolating below the rooting obtained from this study yielded a lower treeline where zone (Sugimoto et al. 2002). Recognising these mutual dry conditions of the basins prevent tree growth, coin- relationships is crucial for understanding the patterns of ciding with the present lower forest boundaries. Exist- forest distribution and tree biomass in the Mongolian ence of forests below the threshold of 160 mm MAP can forest-steppe. However, this causal network alone cannot be explained by additional water supply through lateral explain the present forest-distribution pattern, as it is water fluxes, cumulating in concave positions and toe additionally influenced by other - mostly anthropogenic slopes (Klinge et al. 2021). The PFA moreover suggests - factors that may lead to a discrepancy between the ac- a potential for greater forest areas on south-facing tual and potential forest area (PFA). slopes. This mismatch reconfirms that MAP and MGST alone cannot explain forest distribution, which is a result Potential forest area (PFA) of a more complex causal network as explained in the The relief-limited potential forest area (PFA ) obtained beginning of this chapter. Short growing seasons and from this study suggests a potential for forest expansion, long-lasting snow cover prevent the expansion of forests both downslope towards the basins, and upslope towards into upper valleys and onto the mountain plateaus of the the high-mountain area. Pediments that widely cover the high-mountain area in the south. Another limiting factor toe slopes in the study area generally provide suitable there is the extensive use of alpine meadows as summer geoecological conditions for tree growth, as confirmed pastures. The upper treeline rises from 2400 m a.s.l. in by several existing small forest stands there. the north to 2600 m a.s.l. in the south of the study area. Klinge et al. Forest Ecosystems (2021) 8:55 Page 14 of 20 Fig. 9 Relationships between leaf area index (LAI) and living-tree biomass (left axis, blue dots), and between LAI and NDVI (right axis, red dots) of 24 plots. NDVI = mean NDVI obtained from seven Sentinel 2 images (Fig. 3) Thereby, small treeless areas on the flat summits of the the past 450 years based on tree ring analysis. In agree- northern mountains may result from the so-called “sum- ment with results from the Tuva region in southern Si- mit effect” (Körner 2012), i.e., particularly harsh condi- beria (Ivanova et al. 2010), the authors did not detect an tions near summits, rather than from a true upper increase in fire frequency during the last decades, but treeline. The projected PFAs suggest more large forests fires became more severe due to drier conditions. The and considerably less small, fragmented forest stands in limited contemporaneity of fire events at different sites the steppe-dominated area in the northern part of the pointed to fire-raising by humans (Hessl et al. 2012). On study area, assuming potential forest-dominated area the other hand, human impact may also lead to reduced there. A shift from potential forest-dominated area to destructiveness of fires, as wood gathering and intensive the presently observed steppe-dominated area may have grazing of livestock reduce available fuel for fires been partially triggered by natural factors such as fire, (Umbanhowar et al. 2009; Hessl et al. 2012). windbreak, insect calamities, and drought, but logging Although the most extensive forest fires in this area and forest pasture most likely caused major forest losses occurred already in 2002, forests have not yet re- in this area. Given the permafrost-promoting effect of established in many burnt forest areas. The most exten- large forests, the proposed forest-dominated area sce- sive burnt forest areas are located in the large forests of nario for the northern part of the study area would in- the high-mountain area and in the upper mountains in volve also greater abundance of permafrost in this area. the forest-dominated area. In contrast, only few burnt forest areas occur in the small and fragmented forests of Impact of fire the steppe-dominated area. We conclude that forest The forest area that burnt down between 1986 and 2017 fragmentation in the steppe-dominated area prevents amounts to 12.8% of the total study area. The loss of forest fires from passing over into neighbouring forest living-tree biomass since the last century adds up to stands and keeps fires rather isolated. The decrease of roughly 15 million tons, which represents more than large forests (size class F4) by fire led to an increase in 45% of the former tree biomass. Nyamjav et al. (2007) small, fragmented forest stands (size class F1), represent- stated that 95% of the actual forest destruction was ing remnants of the former large forests. This change in- caused by forest fires, whereas 5% was due to logging. duced loss of permafrost in these areas. Surviving larch The authors reported an increase of fire events in trees in the remaining forest remnants show enhanced Mongolia during the past decades. Goldammer (2002) fructification (Danilin and Tsogt 2014). Their important assumed that most of the fires were caused by human role as nuclei for forest regeneration is demonstrated by activities. Hessl et al. (2012) investigated fire history over numerous seedlings and saplings growing in the direct Klinge et al. Forest Ecosystems (2021) 8:55 Page 15 of 20 Table. 3 Dimensions (km ) and relative portions (%) of different forest categories and landscape units in the study area at present (actual forest area) and in 1986, prior to the large forest fires Actual forest area Burnt Forest distribution in 1986 Steppe forest Forest-size class 1 2 3 4 1 2 3 4 Absolute area of landscape type (km ) Steppe-dominated area Interior 4.74 18.25 21.51 7.29 3.88 5.59 19.65 22.90 7.92 1042.29 Edge 11.07 12.33 7.65 1.69 10.28 16.19 6.92 1.37 Ratio I/E 0.43 1.48 2.81 4.32 0.54 1.21 3.31 5.80 Forest-dominated area Interior 12.66 82.93 144.04 440.35 115.25 13.90 87.52 120.78 606.55 1907.62 Edge 43.55 52.56 44.21 107.34 32.10 42.10 32.12 105.36 Ratio I/E 0.29 1.58 3.26 4.10 0.43 2.08 3.76 5.76 High-mountain area Interior 1.73 7.51 6.40 693.33 3.74 29.70 65.66 491.08 1021.74 Edge 13.06 6.73 2.32 10.70 15.92 20.71 93.56 Ratio I/E 0.13 1.12 2.76 0.35 1.87 3.17 5.25 Alluvial forest 3.98 3.42 1.09 0.05 2.74 2.74 2.74 485.74 Dune 0.01 27.28 Relative proportion of landscape type (%) Steppe-dominated area Interior 0.07 0.29 0.34 0.11 0.06 0.09 0.31 0.36 0.12 16.40 Edge 0.17 0.19 0.12 0.03 0.16 0.25 0.11 0.02 Sum 0.25 0.48 0.46 0.14 0.25 0.56 0.47 0.15 Forest-dominated area Interior 0.20 1.30 2.27 6.93 1.81 0.22 1.38 1.90 9.54 30.02 Edge 0.69 0.83 0.70 1.69 0.51 0.66 0.51 1.66 Sum 0.88 2.13 2.96 8.62 0.72 2.04 2.41 11.20 High-mountain area Interior 0.03 0.12 0.10 10.91 0.06 0.47 1.03 7.73 16.08 Edge 0.21 0.11 0.04 0.17 0.25 0.33 1.47 Sum 0.23 0.22 0.14 0.23 0.72 1.36 9.20 Alluvial forest 0.13 0.001 0.04 0.07 0.02 7.64 Dune 0.43 surrounding of the forest remnants, in the shade of the Living-tree biomass old trees. Thus, a slow but steady re-immigration of In contrast to the close relationships of forest distribu- larch trees into the burnt area proceeds from these for- tion with relief and climate, living-tree biomass showed est remnants. It may take up to 200 years until a forest no significant correlation with topographic parameters. regenerates to its state prior to a fire (Nyamjav et al. It turned out that forests of the Mongolian forest-steppe 2007). have highly variable living-tree biomass. In addition to Fires occur frequently in semi-arid environment (Hessl natural impacts on forests (e.g., fire, windbreak, insect et al. 2012). Thus, L. sibirica is fire-adapted to a certain calamities, and drought), logging and forest pasture may degree. Its survival of a fire depends on the type of fire affect living-tree biomass. Alluvial forests exist where (crown, surface or ground fire), fire intensity, season, river channels hamper wood pasture, logging, and forest and soil moisture. The prevalent survival of forest stands fires. These alluvial forests usually consist of old larch in depressions, erosion channels, and on toe slopes dem- trees and have large tree biomass. Open larch forests on onstrates the importance of soil moisture for tree dunes are also made up by very old trees, but have low survival. stand density and tree biomass. Klinge et al. Forest Ecosystems (2021) 8:55 Page 16 of 20 Table. 4 Total living-tree biomass (10 g) in different forest categories of the study area Actual forest area Forest distribution in 1986 Forest-size class 1 2 3 4 Sum 1 2 3 4 Sum Steppe-dominated area Interior 68,892 398,195 474,487 132,398 1,073,972 81,348 428,835 505,209 143,948 1,159,341 Edge 149,217 191,236 130,499 31,053 502,005 138,668 251,237 118,171 25,152 533,229 Sum 218,109 589,431 604,986 163,450 1,575,976 220,016 680,073 623,380 169,100 1,692,569 Forest-dominated area Interior 277,732 1,809,560 3,177,481 8,000,534 13,265,306 304,883 1,909,714 2,664,408 11,020,218 15,899,224 Edge 718,971 815,498 754,423 1,977,866 4,266,758 530,041 653,229 548,092 1,941,327 3,672,689 Sum 996,703 2,625,058 3,931,904 9,978,400 17,532,064 834,925 2,562,943 3,212,500 12,961,545 19,571,913 High-mountain area Interior 37,963 163,881 141,084 0 342,928 82,118 648,002 1,448,389 8,922,284 11,100,793 Edge 215,657 104,498 39,519 0 359,674 176,719 247,038 353,449 1,723,920 2,501,126 Sum 253,620 268,379 180,603 0 702,602 258,837 895,041 1,801,838 10,646,204 13,601,919 Alluvial forest 350,149 353,095 Total sum 20,160,791 35,219,497 Comparison of tree biomass Sum 1986 1,313,778 4,138,056 5,637,718 23,776,849 34,866,401 Sum 2018 1,468,432 3,482,868 4,717,493 10,141,850 19,810,642 Loss by fire − 154,655 655,189 920,225 13,634,999 15,055,759 Percentage (%) −11.8 15.8 16.3 57.3 43.2 In attempts to assess biomass at landscape scale, the statistically significant correlation between NDVI and NDVI is commonly used as a biomass proxy. However, living-tree biomass. Instead, the NDVI proved to be a its suitability depends on the scale and data resolution. suitable indicator of the growing conditions for the en- Dulamsuren et al. (2016) successfully applied NDVI at tire forest vegetation (Erasmi et al. 2021). regional scale for biomass estimation in Mongolia. How- The differences in mean tree biomass between the forest ever, in our local-scale analysis we did not obtain a categories distinguished in our study were up to 85 Table. 5 Potential forest area (PFA) and living-tree biomass as controlled by climate (PFA ) and relief (PFA ) c r Potential forest area PFA Potential forest area PFA c r Forest-size class 1 2 3 4 Steppe 1 2 3 4 Steppe Area (km ) Interior 11.49 69.21 77.30 2590.66 2692.99 11.17 69.52 91.04 2940.99 2308.48 Edge 35.06 38.45 23.15 322.99 38.46 42.14 28.87 330.62 Ratio I/E 0.33 1.80 3.34 8.02 0.29 1.65 3.15 8.90 Tree biomass (×10 g) Total sum Total sum Interior 251,930 1,510,158 1,705,159 47,068,688 50,535,934 245,059 1,516,857 2,008,287 53,433,781 57,203,985 Edge 578,839 596,605 395,092 5,951,310 7,521,845 635,056 653,769 492,614 6,091,966 7,873,404 Sum 830,768 2,106,762 2100,250 53,019,998 58,057,779 880,115 2,170,625 2500,901 59,525,748 65,077,389 Comparison of tree biomasses (×10 g) PFA - 1986 − 483,009 −2,031,294 −3,537,468 29,243,149 23,191,377 −433,662 −1,967,431 −3,136,817 35,748,898 30,210,988 Percentage (%) −58.1 −96.4 −168.4 55.2 39.9 −49.3 −90.6 −125.4 60.1 46.4 PFA - 2018 −637,664 −1,376,105 −2,617,242 42,878,148 38,247,136 − 588,317 −1,312,242 −2,216,592 49,383,898 45,266,747 Percentage (%) −76.8 −65.3 −124.6 80.9 65.9 −66.8 −60.5 −88.6 83.0 69.6 Klinge et al. Forest Ecosystems (2021) 8:55 Page 17 of 20 − 1 Mg·ha . Our field measurements showed that the least be considered in the PFA projection. For example, large mean tree biomasses occurred in the forest-size class G1 forests (size class F4), as predominantly obtained from − 1 (142 Mg·ha ) of the steppe-dominated area and in the the PFA delineation, are more prone to severe fires than − 1 class F4 (182 Mg·ha ) of the forest-dominated area. The fragmented forest stands. In addition, due to the long- reduced living-tree biomass of the small, fragmented for- lasting human influence, reconstructing the natural pro- ests in the steppe-dominated area (G1) can be explained portion between steppe and forest in this region remains by enhanced forest use, reducing the stand basal area. As a major research challenge (Klinge and Sauer 2019). Hu- therefore solar irradiation on the ground is increased, man impact already started with the extinction of large there is no permafrost under these forests. The reduced herbivores like elephantine, and the reduction of wild living-tree biomass of the largest forests (size class F4) has animal herds since the Mesolithic period. It continued a different reason. There, the permafrost table approaches with the breeding of domestic animals and the develop- the surface and hinders deep rooting of trees. Trees are ment of pasture since the Neolithic period, which started therefore highly prone to windthrow, which explains the around 4.7 ka BP with the Afanasievo culture in the Altai reduced living-tree biomass in this forest category. Mountains (Kovalev and Erdenebaatar 2009). Furthermore, forest edges showed reduced living-tree Tchebakova et al. (2009) modelled potential vegetation biomass. Forest edges represent natural zones of forest changes across Siberia based on climate-change scenar- expansion and retreat (Sommer and Treter 1999), ios projecting warmer and drier climate. The authors whereby temporal climatic variations control these fluc- forecast an increase of forest-steppe and grassland areas. tuations. Forest edges may have a fringe of dead trees at They assume that drier conditions and larger amounts their outer boundary, and their outer boundary may also of fuel due to enhanced tree mortality will lead to an in- be dissected. In addition, logging and pasture is more in- crease in frequency and destructiveness of fires. tensive at the forest edges than in the interiors. Due to the lower tree density, the living-tree biomass is gener- Conclusions ally lower at the forest edges than in the interiors. How- A combination of tree-biomass determination, perma- ever, we found exceptions to this rule in the forest-size frost detection in soil profiles, remote sensing and classes G1 and F4, where the forest interiors and edges climate-data analysis allowed us to identify factors con- had similar tree biomasses. In the class G1, the tree bio- trolling larch-forest distribution and living-tree biomass masses of the small interior forest areas are similarly low in the northern Khangai Mountains, central Mongolia. as those of the forest edges. In the class F4, the forest The identified topographic and climatic thresholds for edges have large tree biomasses compared to all other forest growth enabled us to delineate the potential forest forest edges. This can be explained by the effect of area (PFA), which was much larger than the actual forest permafrost. A shallow permafrost table in the interior area. Forest fires destroyed 43% of the forest area and part of F4 forests causes reduced tree biomass in the for- 45% of the living-tree biomass in the study area over the est interior as described above. Towards the forest edges, period 1986–2017. They mostly affected large forest the depth of the permafrost table increases. There, dur- stands in the upper mountains. Permafrost, which was ing summer permafrost supplies meltwater to the trees widespread under large forests, disappeared soon after that are otherwise close to the climatic threshold of tree the destruction of a large forest stand. growth at the drier forest edges. This effect, together In contrast to forest distribution, living-tree biomass with higher precipitation in the upper mountains may showed no correlation with topographic and climatic pa- also explain the existence of forests on south-facing rameters. We found neither significant differences in slopes in the higher mountains (Fig. 5). living-tree biomass between forests with different fire Interestingly, forest stands that experienced non-lethal history, degree of exploitation, and soil properties, nor fire events or selective logging had similar tree bio- between most forest-size classes. Only forest edges and masses as pristine forests. A possible explanation is that small, fragmented forest stands had significantly less tree moderate thinning of forests may improve the growing biomass than all other forest categories. Neither non- conditions for the remaining trees, as it leads to reduced lethal fires nor selective logging seriously reduced living- competition for water and increased nutrient supply tree biomass. We conclude that these impacts remove from ash, and the melting permafrost leads to a tempor- tree biomass, but also stimulate growth of the remaining ary increase of soil moisture and allows for deeper trees by reducing competition. rooting. Based on relief thresholds for forest growth, we ob- 2 9 The estimated maximum tree biomass of the PFA tained a PFA of 3552 km with 65 × 10 g tree biomass, 9 2 (58–65 × 10 g) was twice the tree biomass in 1986 and based on climatic thresholds a PFA of 3113 km 9 9 (35 × 10 g) and three times the actual tree biomass of with 58 × 10 g tree biomass, corresponding to 323% 20 × 10 g. However, several relevant factors could not and 288% of the actual tree biomass, respectively. Klinge et al. Forest Ecosystems (2021) 8:55 Page 18 of 20 However, these estimates do not consider several rele- Declarations vant factors such as herbivore grazing and plant compe- Ethics approval and consent to participate tition. In addition, long-lasting human impact (at Not applicable. millennial timescale) plays an important role for the vegetation pattern as well, which needs further Consent for publication Not applicable. investigation. Competing interests Abbreviations The authors declare that they have no competing interests. a.s.l.: Above the sea level; dbh: Tree diameter at breast height; DEM: Digital elevation model; LAI: Leaf area index; MAP: Mean annual precipitation; Author details MGS: Mean growing season, may–september; MGST: Mean growing season Department of Physical Geography, Institute of Geography, University of temperature; NDVI: Normalized differentiated vegetation index; PFA: Potential Göttingen, Goldschmidtstraße 5, 37077 Göttingen, Germany. Applied forest area; SE: Standard error; SRTM: Shuttle radar topography mission Vegetation Ecology, Faculty of Environment and Natural Resources, University of Freiburg, Tennenbacher Str. 4, 79106 Freiburg, Germany. Institute of Farm Economics, Thünen Institute, Bundesallee 63, 38116 Supplementary Information 4 Braunschweig, Germany. Department of Biology, School of Arts and The online version contains supplementary material available at https://doi. Sciences, National University of Mongolia, Baga toiruu 47, Sukhbaatar duureg, org/10.1186/s40663-021-00333-9. Ulaanbaatar, Mongolia. Received: 20 April 2021 Accepted: 22 July 2021 Additional file 1: Table S1 Mean living-tree biomass (above and be- lowground) for different forest categories and site conditions. Plots, where site conditions were not clearly identified, were excluded from the respective part of the analysis. The two lower and higher forest-size clas- References ses in the forest-dominated area were combined for the statistical ana- Academy of Sciences of Mongolia, Academy of Sciences of USSR (1990) National lysis, because of the small dataset for forest edges. SE = standard error, Atlas of the Peoples Republic of Mongolia, Ulaanbaatar, Moscow n = number of plots. Underlined data are not representative because of Batima P, Natsagdorj L, Gombluudev P, Erdenetsetseg B (2005) Observed climate insufficient size of the respective dataset. Fig. S1 NDVI of forests in the change in Mongolia. AIACC Working Papers 12:1–25 study area. Arithmetic means of NDVI from seven Sentinel 2 satellite im- Battulga P, Tsogtbaatar J, Dulamsuren Ch, Hauck M (2013) Equations for ages (17.05.2018, 11.06.2018, 25.08.2018, 4.09.2018, 14.09.2018, 19.09.2017, estimating the above-ground biomass of Larix sibirica in the forest-steppe of 16.07.2016). The shaded relief illustration is based on TanDEM-X data. Mongolia. J Forest Res 24(3):431–437. https://doi.org/10.1007/s11676-013-03 75-4 Dagvadorj D, Natsagdorj L, Dorjpurev J, Namkhainyam B (2009) Mongolia Acknowledgements assessment report on climate change 2009, Ulaanbaatar, Mongolia We thank Ms. Daramragchaa Tuya from the Tarvagatai Nuruu National Park Danilin IM (1995) Structure and biomass of larch stands regenerating naturally (Tosontsengel Sum, Zavkhan Aimag, Mongolia) for her invaluable support of after clearcut logging. Water Air Soil Pollut 82:125–131. https://doi.org/10.1 our research. We wish to express our gratitude to our Mongolian colleagues 007/978-94-017-0942-2_14 Mr. Amarbayasgalan, Mr. Enkhjargal, Mr. Enkh-Agar, Ms. Munkhtuya. We Danilin IM, Tsogt Z (2014) Dynamics of structure and biological productivity of greatly appreciated their hospitality and help with the fieldwork. Our thanks post-fire larch forests in the northern Mongolia. Contemp Probl Ecol 7(2): also go to the German students Martine Koob, Tino Peplau, Janin Klaassen 158–169. https://doi.org/10.1134/S1995425514020036 and Tim Rollwage for their great support with the biomass measurements Dashtseren A, Ishikawa M, Iijima Y, Jambaljav Y (2014) Temperature regimes of during the fieldwork in Mongolia. the active layer and seasonally frozen ground under a forest-steppe mosaic, The German Aerospace Centre (DLR) liberally provided the TanDEM-X data Mongolia. Permafrost Periglac Proc 25(4):295–306. https://doi.org/10.1002/ for the study area (DEM_FOREST 1106). The fieldwork in 2014 was funded by ppp.1824 the Volkswagen Foundation in the frame of the project 87175 “Forest regen- DeLuca TH, Boisvenue C (2012) Boreal forest soil carbon: distribution, function eration and biodiversity at the forest-steppe border of the Altay and Khangay and modelling. Forestry 85(2):161–184. https://doi.org/10.1093/forestry/ Mountains under contrasting development of livestock numbers in cps003 Kazakhstan and Mongolia” granted to M. Hauck, Ch. Dulamsuren and C. Dulamsuren Ch, Hauck M (2008) Spatial and seasonal variation of climate on Leuschner. The subsequent work was funded by the Deutsche Forschungs- steppe slopes of the northern Mongolian mountain taiga. Grassl Sci 54(4): gemeinschaft (DFG), project number 385460422 approved to M. Klinge, D. 217–230. https://doi.org/10.1111/j.1744-697X.2008.00128.x Sauer and M. Frechen. Dulamsuren Ch, Hauck M, Bader M, Osokhjargal D, Oyungerel S, Nyambayar S, Runge M, Leuschner C (2009) Water relations and photosynthetic performance in Larix sibirica growing in the forest-steppe ecotone of Authors’ contributions northern Mongolia. Tree Physiol 29(1):99–110. https://doi.org/10.1093/ MK conceived the ideas; MK, ChD, FS, MH, UB and DS participated fieldwork treephys/tpn008 and collected the data; MK, ChD and SE analyzed the data; MK, ChD, MH and Dulamsuren Ch, Hauck M, Khishigjargal M, Leuschner HH, Leuschner C (2010a) DS wrote the paper. The author(s) read and approved the final manuscript. Diverging climate trends in Mongolian taiga forests influence growth and regeneration of Larix sibirica. Oecologia 163(4):1091–1102. https://doi.org/10.1 007/s00442-010-1689-y Funding Dulamsuren Ch, Hauck M, Leuschner C (2010c) Recent drought stress leads to This study was funded by the Volkswagen Foundation (project-no. 871759) growth reductions in Larix sibirica in the western Khentey, Mongolia. Glob and by the German Research Council (Deutsche Forschungsgemeinschaft, Change Biol 16:3024–3035. https://doi.org/10.1111/j.1365-2486.2009.02147.x DFG), (project no. 385460422). Dulamsuren Ch, Hauck M, Leuschner HH, Leuschner C (2010b) Gypsy moth- induced growth decline of Larix sibirica in a forest-steppe ecotone. Availability of data and materials Dendrochronologia 28(4):207–213. https://doi.org/10.1016/j.dendro.2009.05. The datasets used and/or analyzed during the current study are available 007 from the corresponding author on reasonable request. Dulamsuren Ch, Hauck M, Mühlenberg M (2008) Insect and small mammal Clime data are publicly available from the CHELSA data set (https://chelsa- herbivores limit tree establishment in northern Mongolian steppe. Plant Ecol climate.org/). 195(1):143–156. https://doi.org/10.1007/s11258-007-9311-z Klinge et al. Forest Ecosystems (2021) 8:55 Page 19 of 20 Dulamsuren Ch, Khishigjargal M, Leuschner C, Hauck M (2014) Response of tree- Khishigjargal M, Dulamsuren C, Lkhagvadorj D, Leuschner C, Hauck M (2013) ring width to climate warming and selective logging in larch forests of the Contrasting responses of seedling and sapling densities to livestock density Mongolian Altai. J Plant Ecol 7(1):24–38. https://doi.org/10.1093/jpe/rtt019 in the Mongolian forest-steppe. Plant Ecol 214(11):1391–1403. https://doi. Dulamsuren Ch, Klinge M, Bat-Enerel B, Ariunbaatar T, Tuya D (2019) Effects of org/10.1007/s11258-013-0259-x forest fragmentation on organic carbon pool densities in the Mongolian Klinge M, Böhner J, Erasmi S (2015) Modeling forest lines and forest distribution forest-steppe. Forest Ecol Manag 433:780–788. https://doi.org/10.1016/j. patterns with remote-sensing data in a mountainous region of semiarid foreco.2018.10.054 Central Asia. Biogeosciences 12(10):2893–2905. https://doi.org/10.5194/ bg-12-2893-2015 Dulamsuren Ch, Klinge M, Degener J, Khishigjargal M, Chenlemuge T, Bat-Enerel B, Yeruult Y, Saindovdon D, Ganbaatar K, Tsogtbaatar J, Leuschner C, Hauck Klinge M, Dulamsuren C, Erasmi S, Karger DN, Hauck M (2018) Climate effects on M (2016) Carbon pool densities and a first estimate of the total carbon pool vegetation vitality at the treeline of boreal forests of Mongolia. in the Mongolian forest-steppe. Glob Change Biol 22(2):830–844. https://doi. Biogeosciences 15(5):1319–1333. https://doi.org/10.5194/bg-15-1319-2018 org/10.1111/gcb.13127 Klinge M, Sauer D (2019) Spatial pattern of late glacial and Holocene climatic and Erasmi S, Klinge M, Dulamsuren Ch, Schneider F, Hauck M (2021) Modelling the environmental development in Western Mongolia - a critical review and productivity of Siberian larch forests from Landsat NDVI time series in synthesis. Quaternary Sci Rev 210:26–50. https://doi.org/10.1016/j.quascirev.2 fragmented forest stands of the Mongolian forest-steppe. Environ Monit 019.02.020 Assess 193(4):200. https://doi.org/10.1007/s10661-021-08996-1 Klinge M, Schneider F, Dulamsuren Ch, Arndt K, Bayarsaikhan U, Sauer D (2021) Feldmann E, Glatthorn J, Hauck M, Leuschner C (2018) A novel empirical Interrelations between relief, vegetation, disturbances, and permafrost in the approach for determining the extension of forest development stages in forest-steppe of Central Mongolia. Earth Surf Process Landforms 46(9):1766– temperate old-growth forests. Eur J Forest Res 137(3):321–335. https://doi. 1782. https://doi.org/10.1002/esp.5116 org/10.1007/s10342-018-1105-4 Körner C (2012) Alpine treelines: functional ecology of the global high elevation Goldammer G (2002) Fire Situation in Mongolia. International Forest Fire News tree limits. Springer, Basel, Heidelberg, New York, Dordrecht, London 26:75–83 Kovalev AA, Erdenebaatar D (2009) Discovery of new cultures of the bronze age Goldammer G (ed) (2007) International Forest Fire News. UNECE and FAO 36 in Mongolia according to the data obtained by the international central Goodale CL, Apps MJ, Birdsey RA, Field CB, Heath LS, Houghton RA, Jenkins JC, Asian archaeological expedition. In: Bemmann J, Parzinger H, Pohl E, Kohlmaier GH, Kurz W, Liu S, Nabuurs G-J, Nilsson S, Shvidenko AZ (2002) Tseveendorj D (eds) Current archaeological research in Mongolia, Bonn, pp. Forest carbon sinks in the northern hemisphere. Ecol Appl 12(3):891–899. 149–170 https://doi.org/10.1890/1051-0761(2002)012[0891:FCSITN]2.0.CO;2 Kowalkowski K, Starkel L (1984) Altitudinal belts of geomorphic pro¬cesses in the Hais M, Chytrý M, Horsák M (2016) Exposure-related forest-steppe: a diverse southern Changai Mts. (Mongolia). Stud Geomorphol Carpatho-Balcan 18:95– landscape type determined by topography and climate. J Arid Environ 135: 115 75–84. https://doi.org/10.1016/j.jaridenv.2016.08.011 Liu H, Park Williams A, Allen CD, Guo D, Wu X, Anenkhonov OA, Liang E, Sandanov DV, Yin Y, Qi Z, Badmaeva NK (2013) Rapid warming accelerates Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, Tyukavina A, Thau tree growth decline in semi-arid forests of inner Asia. Glob Change Biol 19(8): D, Stehman SV, Goetz SJ, Loveland TR, Kommareddy A, Egorov A, Chini L, st 2500–2510. https://doi.org/10.1111/gcb.12217 Justice CO, Townshend JRG (2013) High-resolution global maps of 21 - century forest cover change. Science 342(6160):850–853. https://doi.org/1 Lkhagvadorj D, Hauck M, Dulamsuren Ch, Tsogtbaatar J (2013a) Pastoral 0.1126/science.1244693 nomadism in the forest-steppe of the Mongolian Altai under a changing Hauck M, Dulamsuren Ch, Bayartogtokh B, Ulykpan K, Burkitbaeva UD, economy and a warming climate. J Arid Environ 88:82–89. https://doi.org/1 Otgonjargal E, Titov SV, Enkhbayar T, Sundetpaev AK, Beket U, Leuschner C 0.1016/j.jaridenv.2012.07.019 (2014) Relationships between the diversity patterns of vascular plants, lichens Lkhagvadorj D, Hauck M, Dulamsuren Ch, Tsogtbaatar J (2013b) Twenty years and invertebrates in the central Asian forest-steppe ecotone. Biodivers after decollectivization: Mobile livestock husbandry and its ecological impact Conserv 23(5):1105–1117. https://doi.org/10.1007/s10531-014-0648-z in the Mongolian forest-steppe. Hum Ecol 41(5):725–735. https://doi.org/10.1 Hessl AE, Ariya U, Brown P, Byambasuren O, Green T, Jacoby G, Sutherland EK, 007/s10745-013-9599-3 Nachin B, Maxwell RS, Pederson N, de Grandpré L, Saladyga T, Tardif JC Luyssaert S, Inglima I, Jung M, Richardson AD, Reichstein M, Papale D, Piao S, (2012) Reconstructing fire history in Central Mongolia from tree-rings. Int J Schulze E-D, Wingate L, Matteucci G, Aragao L, Aubinet M, Beer C, Bernhofer Wildland Fire 21(1):86. https://doi.org/10.1071/WF10108 C, Black K, Bonal D, Bonnefond J, Chambers J, Ciais P, Cook B, Davis K, Dolman A, Gielen B, Goulden M, Grace J, Granier A, Grelle A, Griffis T, Hilbig W (1995) The vegetation of Mongolia. SPB Acad Publ, Amsterdam Grünwald T, Guidolotti G, Hanson A, Harding R, Hollinger D, Hutyra L, Kolari IPCC (2013) Climate change 2013 - the physical science basis. Cambridge P, Kruijt B, Kutsch W, Lagergren F, Laurila T, Law B, Le Maire G, Lindroth A, University Press, Cambridge Loustau D, Malhi Y, Mateus J, Migliavacca M, Misson L, Montagnani L, Ivanova GA, Ivanov VA, Kukavskaya EA, Soja AJ (2010) The frequency of forest Moncrieff J, Moors E, Munger W, Nikinmaa E, Ollinger S, Pita G, Rebmann C, fires in scots pine stands of Tuva, Russia. Environ Res Lett 5(1):15002. https:// Roupsard O, Saigusa N, Sanz M, Seufert G, Sierra C, Smith M, Tang J, Valentini doi.org/10.1088/1748-9326/5/1/015002 R, Vesala T, Janssens I (2007) CO balance of boreal, temperate, and tropical Jacob M, Bade C, Calvete H, Dittrich S, Leuschner C, Hauck M (2013) Significance 2 forests derived from a global database. Glob Change Biol 13(12):2509–2537. of over-mature and decaying trees for carbon stocks in a central European https://doi.org/10.1111/j.1365-2486.2007.01439.x natural spruce forest. Ecosystems 16(2):336–346. https://doi.org/10.1007/s1 0021-012-9617-0 Miehe G, Miehe S, Böhner J, Kaiser K, Hensen I, Madsen D, Liu J, Opgenoorth L (2014) How old is the human footprint in the world’s largest alpine Jarvis P, Saugier B, Schulze E-D (2001) Productivity of boreal forests. In: Roy J, ecosystem? A review of multiproxy records from the Tibetan plateau from Saugier B, Mooney H (eds) Terrestrial global productivity. Academic Press, the ecologists’ viewpoint. Quaternary Sci Rev 86:190–209. https://doi.org/10.1 San Diego, pp 211–244. https://doi.org/10.1016/B978-012505290-0/50011-9 016/j.quascirev.2013.12.004 Karger DN, Conrad O, Böhner J, Kawohl T, Kreft H, Soria-Auza RW, Zimmermann Miehe G, Miehe S, Will M, Opgenoorth L, Duo L, Dorgeh T, Liu J (2007) An NE, Linder HP, Kessler M (2017) Climatologies at high resolution for the inventory of forest relicts in the pastures of southern Tibet (Xizang a.R., earth’s land surface areas. Sci Data 4(1):170122. https://doi.org/10.1038/sda China). Plant Ecol 194(2):157–177. https://doi.org/10.1007/s11258-007-9282-0 ta.2017.122 Mukhortova L, Schepaschenko D, Shvidenko A, McCallum I, Kraxner F (2015) Soil Khansaritoreh E, Dulamsuren Ch, Klinge M, Ariunbaatar T, Bat-Enerel B, Batsaikhan contribution to carbon budget of Russian forests. Agric For Meteorol 200:97– G, Ganbaatar K, Saindovdon D, Yeruult Y, Tsogtbaatar J, Tuya D, Leuschner C, 108. https://doi.org/10.1016/j.agrformet.2014.09.017 Hauck M (2017b) Higher climate warming sensitivity of Siberian larch in small Nyamjav B, Goldammer G, Uibrig H (2007) Fire situation in Mongolia. than large forest islands in the fragmented Mongolian forest steppe. Glob International Forest Fire News 36:46–66 Change Biol 23(9):3675–3689. https://doi.org/10.1111/gcb.13750 Pan Y, Birdsey RA, Fang J, Houghton R, Kauppi PE, Kurz WA, Phillips OL, Khansaritoreh E, Eldarov M, Ganbaatar K, Saindovdon D, Leuschner C, Hauck M, Shvidenko A, Lewis SL, Canadell JG, Ciais P, Jackson RB, Pacala SW, McGuire Dulamsuren C (2017a) Age structure and trends in annual stem increment of AD, Piao S, Rautiainen A, Sitch S, Hayes D (2011) A large and persistent Larix sibirica in two neighboring Mongolian forest–steppe regions differing carbon sink in the world’s forests. Science 333(6045):988–993. https://doi. in land use history. Trees 31(6):1973–1986. https://doi.org/10.1007/s00468-01 org/10.1126/science.1201609 7-1601-z Klinge et al. Forest Ecosystems (2021) 8:55 Page 20 of 20 Poulter B, Pederson N, Liu H, Zhu Z, D’Arrigo R, Ciais P, Davi N, Frank D, Leland C, Myneni R, Piao S, Wang T (2013) Recent trends in inner Asian forest dynamics to temperature and precipitation indicate high sensitivity to climate change. Agric For Meteorol 178-179:31–45. https://doi.org/10.1016/j.a grformet.2012.12.006 Richter H, Haase G, Barthel H (1963) Die Golezterrassen. Petermanns Geogr Mitt 107:183–192 Sakamoto K, Tomonari M, Ariya U, Nakagiri E, Matsumoto TK, Akaji Y, Otoda T, Hirobe M, Nachin B (2021) Effects of large-scale forest fire followed by illegal logging on the regeneration of boreal forests in Mongolia. Landscape Ecol Eng 17(3):267–279. https://doi.org/10.1007/s11355-021-00457-8 Schlütz F, Dulamsuren Ch, Wieckowska M, Mühlenberg M, Hauck M (2008) Late Holocene vegetation history suggests natural origin of steppes in the northern Mongolian mountain taiga. Palaeogeogr Palaeoclimatol Palaeoecol 261(3-4):203–217. https://doi.org/10.1016/j.palaeo.2007.12.012 Shvidenko AZ, Schepaschenko D (2014) Carbon balance of Russian forests. Siber Forest J 1:69–92 Sommer M, Treter U (1999) Die Lärchenwälder der Gebirgswaldsteppe in den Randgebieten des Uvs Nuur- Beckens. Erde 130:173–188 Sugimoto A, Yanagisawa N, Naito D, Fujita N, Maximov TC (2002) Importance of permafrost as a source of water for plants in east Siberian taiga. Ecol Res 17(4):493–503. https://doi.org/10.1046/j.1440-1703.2002.00506.x Tchebakova NM, Parfenova E, Soja AJ (2009) The effects of climate, permafrost and fire on vegetation change in Siberia in a changing climate. Environ Res Lett 4(4):45013. https://doi.org/10.1088/1748-9326/4/4/045013 Thurner M, Beer C, Santoro M, Carvalhais N, Wutzler T, Schepaschenko D, Shvidenko A, Kompter E, Ahrens B, Levick SR, Schmullius C (2014) Carbon stock and density of northern boreal and temperate forests. Glob Ecol Biogeogr 23(3):297–310. https://doi.org/10.1111/geb.12125 Treter U (1996) Gebirgs-Waldsteppe in der Mongolei. Geogr Rundsch 48:655–661 Tsogtbaatar J (2004) Deforestation and reforestation needs in Mongolia. Forest Ecol Manag 201(1):57–63. https://doi.org/10.1016/j.foreco.2004.06.011 Umbanhowar CE, Shinneman AL, Tserenkhand G, Hammon ER, Lor P, Nail K (2009) Regional fire history based on charcoal analysis of sediments from nine lakes in western Mongolia. Holocene 19(4):611–624. https://doi.org/1 0.1177/0959683609104039 Unkelbach J, Dulamsuren C, Punsalpaamuu G, Saindovdon D, Behling H (2017) Late Holocene vegetation, climate, human and fire history of the forest- steppe-ecosystem inferred from core G2-a in the ‘Altai Tavan Bogd’ conservation area in Mongolia. Veget Hist Archaeobot 27(5):665–677. https:// doi.org/10.1007/s00334-017-0664-5 Unkelbach J, Kashima K, Enters D, Dulamsuren C, Punsalpaamuu G, Behling H (2019) Late Holocene (Meghalayan) palaeoenvironmental evolution inferred from multi-proxy-studies of lacustrine sediments from the Dayan Nuur region of Mongolia. Palaeogeogr Palaeoclimatol Palaeoecol 530:1–14. https:// doi.org/10.1016/j.palaeo.2019.05.021 Yang Y, Wang Z, Li J, Gang C, Zhang Y, Zhang Y, Odeh I, Qi J (2016) Comparative assessment of grassland degradation dynamics in response to climate variation and human activities in China, Mongolia, Pakistan and Uzbekistan from 2000 to 2013. J Arid Environ 135:164–172. https://doi.org/10.1016/j.ja ridenv.2016.09.004 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png "Forest Ecosystems" Springer Journals http://www.deepdyve.com/lp/springer-journals/geoecological-parameters-indicate-discrepancies-between-potential-and-ajoBhDKBQj

Loading next page...

References (40)

E Feldmann, J Glatthorn, M Hauck, C Leuschner (2018)
A novel empirical approach for determining the extension of forest development stages in temperate old-growth forests
Eur J Forest Res, 137
J Unkelbach, C Dulamsuren, G Punsalpaamuu, D Saindovdon, H Behling (2017)
Late Holocene vegetation, climate, human and fire history of the forest-steppe-ecosystem inferred from core G2-a in the ‘Altai Tavan Bogd’ conservation area in Mongolia
Veget Hist Archaeobot, 27
A Dashtseren, M Ishikawa, Y Iijima, Y Jambaljav (2014)
Temperature regimes of the active layer and seasonally frozen ground under a forest-steppe mosaic, Mongolia
Permafrost Periglac Proc, 25
H Richter, G Haase, H Barthel (1963)
Die Golezterrassen
Petermanns Geogr Mitt, 107
IM Danilin, Z Tsogt (2014)
Dynamics of structure and biological productivity of post-fire larch forests in the northern Mongolia
Contemp Probl Ecol, 7
K Sakamoto, M Tomonari, U Ariya, E Nakagiri, TK Matsumoto, Y Akaji, T Otoda, M Hirobe, B Nachin (2021)
Effects of large-scale forest fire followed by illegal logging on the regeneration of boreal forests in Mongolia
Landscape Ecol Eng, 17
P Jarvis, B Saugier, E-D Schulze, J Roy, B Saugier, H Mooney (2001)
Productivity of boreal forests
Terrestrial global productivity
GA Ivanova, VA Ivanov, EA Kukavskaya, AJ Soja (2010)
The frequency of forest fires in scots pine stands of Tuva, Russia
Environ Res Lett, 5
M Klinge, D Sauer (2019)
Spatial pattern of late glacial and Holocene climatic and environmental development in Western Mongolia - a critical review and synthesis
Quaternary Sci Rev, 210
E Khansaritoreh, M Eldarov, K Ganbaatar, D Saindovdon, C Leuschner, M Hauck, C Dulamsuren (2017)
Age structure and trends in annual stem increment of Larix sibirica in two neighboring Mongolian forest–steppe regions differing in land use history
Trees, 31
NM Tchebakova, E Parfenova, AJ Soja (2009)
The effects of climate, permafrost and fire on vegetation change in Siberia in a changing climate
Environ Res Lett, 4
CE Umbanhowar, AL Shinneman, G Tserenkhand, ER Hammon, P Lor, K Nail (2009)
Regional fire history based on charcoal analysis of sediments from nine lakes in western Mongolia
Holocene, 19
IM Danilin (1995)
Structure and biomass of larch stands regenerating naturally after clearcut logging
Water Air Soil Pollut, 82
M Klinge, J Böhner, S Erasmi (2015)
Modeling forest lines and forest distribution patterns with remote-sensing data in a mountainous region of semiarid Central Asia
Biogeosciences, 12
P Batima, L Natsagdorj, P Gombluudev, B Erdenetsetseg (2005)
Observed climate change in Mongolia
AIACC Working Papers, 12
MC Hansen, PV Potapov, R Moore, M Hancher, SA Turubanova, A Tyukavina, D Thau, SV Stehman, SJ Goetz, TR Loveland, A Kommareddy, A Egorov, L Chini, CO Justice, JRG Townshend (2013)
High-resolution global maps of 21st-century forest cover change
Science, 342
M Thurner, C Beer, M Santoro, N Carvalhais, T Wutzler, D Schepaschenko, A Shvidenko, E Kompter, B Ahrens, SR Levick, C Schmullius (2014)
Carbon stock and density of northern boreal and temperate forests
Glob Ecol Biogeogr, 23
L Mukhortova, D Schepaschenko, A Shvidenko, I McCallum, F Kraxner (2015)
Soil contribution to carbon budget of Russian forests
Agric For Meteorol, 200
M Hais, M Chytrý, M Horsák (2016)
Exposure-related forest-steppe: a diverse landscape type determined by topography and climate
J Arid Environ, 135
M Jacob, C Bade, H Calvete, S Dittrich, C Leuschner, M Hauck (2013)
Significance of over-mature and decaying trees for carbon stocks in a central European natural spruce forest
Ecosystems, 16
B Nyamjav, G Goldammer, H Uibrig (2007)
Fire situation in Mongolia
International Forest Fire News, 36
K Kowalkowski, L Starkel (1984)
Altitudinal belts of geomorphic pro¬cesses in the southern Changai Mts. (Mongolia)
Stud Geomorphol Carpatho-Balcan, 18
J Tsogtbaatar (2004)
Deforestation and reforestation needs in Mongolia
Forest Ecol Manag, 201
A Sugimoto, N Yanagisawa, D Naito, N Fujita, TC Maximov (2002)
Importance of permafrost as a source of water for plants in east Siberian taiga
Ecol Res, 17
AE Hessl, U Ariya, P Brown, O Byambasuren, T Green, G Jacoby, EK Sutherland, B Nachin, RS Maxwell, N Pederson, L de Grandpré, T Saladyga, JC Tardif (2012)
Reconstructing fire history in Central Mongolia from tree-rings
Int J Wildland Fire, 21
G Miehe, S Miehe, J Böhner, K Kaiser, I Hensen, D Madsen, J Liu, L Opgenoorth (2014)
How old is the human footprint in the world’s largest alpine ecosystem? A review of multiproxy records from the Tibetan plateau from the ecologists’ viewpoint
Quaternary Sci Rev, 86
G Miehe, S Miehe, M Will, L Opgenoorth, L Duo, T Dorgeh, J Liu (2007)
An inventory of forest relicts in the pastures of southern Tibet (Xizang a.R., China)
Plant Ecol, 194
J Unkelbach, K Kashima, D Enters, C Dulamsuren, G Punsalpaamuu, H Behling (2019)
Late Holocene (Meghalayan) palaeoenvironmental evolution inferred from multi-proxy-studies of lacustrine sediments from the Dayan Nuur region of Mongolia
Palaeogeogr Palaeoclimatol Palaeoecol, 530
Y Pan, RA Birdsey, J Fang, R Houghton, PE Kauppi, WA Kurz, OL Phillips, A Shvidenko, SL Lewis, JG Canadell, P Ciais, RB Jackson, SW Pacala, AD McGuire, S Piao, A Rautiainen, S Sitch, D Hayes (2011)
A large and persistent carbon sink in the world’s forests
Science, 333
M Khishigjargal, C Dulamsuren, D Lkhagvadorj, C Leuschner, M Hauck (2013)
Contrasting responses of seedling and sapling densities to livestock density in the Mongolian forest-steppe
Plant Ecol, 214
H Liu, A Park Williams, CD Allen, D Guo, X Wu, OA Anenkhonov, E Liang, DV Sandanov, Y Yin, Z Qi, NK Badmaeva (2013)
Rapid warming accelerates tree growth decline in semi-arid forests of inner Asia
Glob Change Biol, 19
U Treter (1996)
Gebirgs-Waldsteppe in der Mongolei
Geogr Rundsch, 48
S Luyssaert, I Inglima, M Jung, AD Richardson, M Reichstein, D Papale, S Piao, E-D Schulze, L Wingate, G Matteucci, L Aragao, M Aubinet, C Beer, C Bernhofer, K Black, D Bonal, J Bonnefond, J Chambers, P Ciais, B Cook, K Davis, A Dolman, B Gielen, M Goulden, J Grace, A Granier, A Grelle, T Griffis, T Grünwald, G Guidolotti, A Hanson, R Harding, D Hollinger, L Hutyra, P Kolari, B Kruijt, W Kutsch, F Lagergren, T Laurila, B Law, G Le Maire, A Lindroth, D Loustau, Y Malhi, J Mateus, M Migliavacca, L Misson, L Montagnani, J Moncrieff, E Moors, W Munger, E Nikinmaa, S Ollinger, G Pita, C Rebmann, O Roupsard, N Saigusa, M Sanz, G Seufert, C Sierra, M Smith, J Tang, R Valentini, T Vesala, I Janssens (2007)
CO2 balance of boreal, temperate, and tropical forests derived from a global database
Glob Change Biol, 13
M Sommer, U Treter (1999)
Die Lärchenwälder der Gebirgswaldsteppe in den Randgebieten des Uvs Nuur- Beckens
Erde, 130
CL Goodale, MJ Apps, RA Birdsey, CB Field, LS Heath, RA Houghton, JC Jenkins, GH Kohlmaier, W Kurz, S Liu, G-J Nabuurs, S Nilsson, AZ Shvidenko (2002)
Forest carbon sinks in the northern hemisphere
Ecol Appl, 12
B Poulter, N Pederson, H Liu, Z Zhu, R D’Arrigo, P Ciais, N Davi, D Frank, C Leland, R Myneni, S Piao, T Wang (2013)
Recent trends in inner Asian forest dynamics to temperature and precipitation indicate high sensitivity to climate change
Agric For Meteorol, 178-179
M Klinge, C Dulamsuren, S Erasmi, DN Karger, M Hauck (2018)
Climate effects on vegetation vitality at the treeline of boreal forests of Mongolia
Biogeosciences, 15
AZ Shvidenko, D Schepaschenko (2014)
Carbon balance of Russian forests
Siber Forest J, 1
DN Karger, O Conrad, J Böhner, T Kawohl, H Kreft, RW Soria-Auza, NE Zimmermann, HP Linder, M Kessler (2017)
Climatologies at high resolution for the earth’s land surface areas
Sci Data, 4
Y Yang, Z Wang, J Li, C Gang, Y Zhang, Y Zhang, I Odeh, J Qi (2016)
Comparative assessment of grassland degradation dynamics in response to climate variation and human activities in China, Mongolia, Pakistan and Uzbekistan from 2000 to 2013
J Arid Environ, 135

Publisher: Springer Journals
Copyright: Copyright © The Author(s) 2021
eISSN: 2197-5620
DOI: 10.1186/s40663-021-00333-9
Publisher site: See Article on Publisher Site

Abstract

Background: Forest distribution in the forest-steppe of Mongolia depends on relief, permafrost, and climate, and is highly sensitive to climate change and anthropogenic disturbance. Forest fires and logging decreased the forest area in the forest-steppe of Mongolia. The intention of this study was to identify the geoecological parameters that control forest distribution and living-tree biomass in this semi-arid environment. Based on these parameters, we aimed to delineate the area that forest might potentially occupy and to analyse the spatial patterns of actual and potential tree biomass. Methods: We used a combination of various geographic methods in conjunction with statistical analyses to identify the key parameters controlling forest distribution. In several field campaigns, we mapped tree biomass and ecological parameters in a study area within the Tarvagatai Nuruu National Park (central Mongolia). Forest areas, topographic parameters and vegetation indices were obtained from remote sensing data. Significant correlations between forest distribution and living-tree biomass on one hand, and topographic parameters, climate data, and environmental conditions on the other hand, were used to delineate the area of potential forest distribution and to estimate total living-tree biomass for this area. Results: Presence of forest on slopes was controlled by the factors elevation, aspect, slope, mean annual precipitation, and mean growing-season temperature. Combining these factors allowed for estimation of potential forest area but was less suitable for tree-biomass delineation. No significant differences in mean living-tree biomass existed between sites exposed to different local conditions with respect to forest fire, exploitation, and soil properties. Tree biomass was reduced at forest edges (defined as 30 m wide belt), in small fragmented and in large 9 2 forest stands. Tree biomass in the study area was 20 × 10 g (1,086 km forest area), whereas the potential tree 9 2 biomass would reach up to 65 × 10 g (> 3168 km ). * Correspondence: mklinge1@gwdg.de Department of Physical Geography, Institute of Geography, University of Göttingen, Goldschmidtstraße 5, 37077 Göttingen, Germany Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Klinge et al. Forest Ecosystems (2021) 8:55 Page 2 of 20 Conclusions: The obtained projection suggests that the potential forest area and tree biomass under the present climatic and geoecological conditions is three times that of the present forest area and biomass. Forest fires, which mostly affected large forest stands in the upper mountains, destroyed 43% of the forest area and 45% of the living- tree biomass in the study area over the period 1986–2017. Keywords: Biomass, Fire, Forest-steppe, Geoecological factors, Mongolia, Permafrost Background vicinity of grasslands contributed to forest degradation The Mongolian forest-steppe represents the transition and decrease of forest area (Tsogtbaatar 2004; Dulamsu- zone between the southern limit of the boreal forest in ren et al. 2014). Widespread logging activities after forest Central Asia and the dry region of the Gobi Desert. It is fires with the target to extract the remaining wood from characterized by semi-arid climate and is highly vulner- the forest are thought to delay reforestation due to able to climate change and land-use intensification mechanical damage to young trees (Sakamoto et al. (Poulter et al. 2013; Yang et al. 2016; Khansaritoreh 2021). et al. 2017b). Various ecological stress factors have re- Boreal forests represent an important organic carbon cently reduced Mongolia’s forest area and thus most pool and are thus important for the global climate likely tree biomass as well (Dulamsuren et al. 2010a, (Goodale et al. 2002; Pan et al. 2011). Although most of 2010b; Hansen et al. 2013). Mongolia’s boreal forests the organic carbon in the boreal zone is stored in soils harbor a unique biodiversity of forest species, which is (DeLuca and Boisvenue 2012; Shvidenko and Sche- regionally obliterated due to deforestation (Hauck et al. paschenko 2014; Mukhortova et al. 2015), a considerable 2014). Moreover, conversion of boreal forests to steppe amount of carbon is also stored in the living-tree bio- grassland is estimated to reduce the organic carbon mass. Typically, carbon stocks in the total tree biomass − 1 stock density by roughly 40% not only due to the re- of boreal forests amount to 40–80 Mg C∙ha (Jarvis moval of biomass, but also as the result of carbon losses et al. 2001; Luyssaert et al. 2007; Thurner et al. 2014). from the organic layer (Dulamsuren et al. 2016). Investigations on tree biomass in the Mongolian forest- Drought stress and resulting declines in wood produc- steppe have been carried out in the Altai Mountains, tion and forest regeneration were repeatedly reported, southern Khangai Mountains (Dulamsuren et al. 2016), especially for Siberian larch (L. sibirica Ledeb.), which northern Khangai Mountains (Dulamsuren et al. 2019), makes up approximately 80% of the total forest area and Khentei Mountains (Danilin 1995; Danilin and (Dulamsuren et al. 2009, 2010a, 2010c; Liu et al. 2013). Tsogt 2014). Obtained tree biomass values were in the − 1 Mongolia’s mean annual air temperature has increased range of 123–397 Mg∙ha and thus (due to Mongolia’s by 0.27 K per decade (in total 1.7 K) from 1940 to 2001 position in the southernmost boreal zone) in and beyond (Batima et al. 2005), which is clearly above the global the higher range typical for boreal forests Altogether, average of 0.12 K per decade from 1951 to 2012 (IPCC these studies point to a decrease of average tree biomass 2013). For the period 1940–2006, Dagvadorj et al. (2009) from the more humid north to the drier south of the reported a seasonal differentiated temperature increase Mongolian forest-steppe. At local scale, tree biomass in of 0.11 K per decade in summer and of 0.51 K per dec- the interior of L. sibirica forests exceeds that at the for- ade in winter. For the same period, the authors stated est edges (Dulamsuren et al. 2016). No consistent signifi- spatially varying trends of increasing and decreasing cant differences in tree biomass were found between precipitation. forest stands of varying sizes and between forests grow- In addition, devastating forest fires disturbed large for- ing in grassland- and forest-dominated areas of the est areas in Mongolia over the past decades (Goldammer forest-steppe (Dulamsuren et al. 2019). 2002; Hansen et al. 2013; Nyamjav et al. 2007). Local au- Logging, other kinds of forest use such as forest pas- thorities of the Tarvagatai Nuruu National Park in our ture, and fire-setting have reduced the forest area and study region stated that forest fires became more fre- tree biomass in Central Asia since prehistoric times quent since the 1990s, whereby the most severe fires oc- (Miehe et al. 2007, 2014; Unkelbach et al. 2017, 2019). curred in 1996 and 2002. Goldammer (2002) reported The impact of these activities can be evaluated by esti- that fire fighters with an air fleet were installed from mating the potential extent of forest area based on cli- 1969 until the 1990s, when financial support from Russia matic and topographic parameters (Klinge et al. 2015). ended. Thus, the extensive fire events in 1996 and 2002 The parameters precipitation, temperature and evapor- could not be fought as effectively as previous fires. Fur- ation control the spatial pattern of forest and steppe dis- thermore, a lack in systematic forest management, insuf- tribution in the semi-arid forest-steppe (Nyamjav et al. ficient control of logging and forest pasture in the 2007; Dulamsuren and Hauck 2008; Klinge et al. 2018). Klinge et al. Forest Ecosystems (2021) 8:55 Page 3 of 20 In addition, topographic position plays an important role, as The region has continental climate with cold semi-arid forests are generally limited to north-facing slopes (Klinge conditions (Fig. 2). The monthly mean temperatures at et al. 2015;Haisetal. 2016). Thus, relief is an important fac- Tosontsengel range between − 31.7 °C in January and tor for the existence, vigour and tree density of forests. In 14.7 °C in July. Most of the annual precipitation occurs addition to natural factors, the present forest distribution is during summer, from low-pressure cells brought by the strongly influenced by human impact that increased since westerlies (Batima et al. 2005). In contrast, the Siberian prehistoric times. Logging is done in an unsystematic man- High during winter causes mostly dry conditions. The ner for timber and fuelwood and pervasive in the forest- cold climate promotes discontinuous permafrost, with steppe. Its intensity has often increased after the transition permafrost mainly occurring in valley bottoms, upper from planned to market economy in the 1990s (Dulamsuren mountains, and partially on slopes. The existence of et al. 2014). Livestock kept by pastoral nomads influences permafrost ice requires some soil moisture, whereas dry forest regeneration at forest margin and in the interior of soil conditions lead to dry permafrost, i.e., perennially small forests. In the Mongolian Altai, increased livestock frozen ground without ice. densities promoted the establishment of L. sibirica seedlings The maximum altitudes of the study area of up to due to the creation of gaps in the ground vegetation, but later 3200 m a.s.l. occur in its southern part. They are charac- reduced the density of tree regeneration in the sapling stage, terised by mountain plateaus with cryoplanation terraces as the seedlings are a preferred diet of goats (Khishigjargal (Richter et al. 1963; Kowalkowski and Starkel 1984). et al. 2013). Goat numbers in Mongolia have multiplied since These highest regions above the upper treeline at the 1990s owing to the high economic significance of cash- approx. 2500 m a.s.l. belong to the periglacial belt, with mere wool for the herder households (Lkhagvadorj et al. alpine vegetation and bare, rock-debris covered land sur- 2013a, 2013b). Relative to livestock, browsing by wild ungu- faces (Klinge et al. 2018). In the northern part, the lates is of subordinate importance due to lower densities and mountains are lower, and mountain forest-steppe covers a large hunting pressure. The regeneration success of L. sibir- the north-facing slopes up to the summits. The main ica is primarily dependant on moisture availability and her- valleys run from south to north, leading into the east- bivory (Dulamsuren et al. 2008; Khishigjargal et al. 2013). It west running valley of the Ider Gol (Gol: Mongolian for is only loosely related to fire in the Mongolian forest steppe, River) at an elevation of 1600 m a.s.l.. The geological as the dry climate (supported by livestock) generates gaps in basement consists of Permian metamorphosed sedi- the ground vegetation where seedlings can establish (Danilin mentary and acid plutonic rock, and Carboniferous 1995;Dulamsurenet al. 2010b). Tree-ring chronologies often mafic rock (Academy of Sciences of Mongolia, Academy show annual tree establishment over longer periods with of Sciences of USSR, 1990). Coarse detritus of these bed- moist climate, but unrelated to fire (Dulamsuren et al. 2010a, rocks forms slope debris, which is often mixed with and 2010c; Khansaritoreh et al. 2017a). covered by sandy to silty aeolian deposits. Based on the state of knowledge described above, we Dense, extensive forests occur south of the Ider Gol, addressed the following hypotheses: whereas north of the river, forests are more fragmented and steppe vegetation is dominant (Dulamsuren et al. (I) Climatic and topographic parameters limiting the 2019). A clear spatial pattern of forests (made up of L. general distribution of larch forests in the study sibirica) on north-facing slopes and steppe on south- area can be deduced by spatial analysis of remote facing slopes is typical in the forest-steppe of Mongolia sensing data. (Hilbig 1995; Treter 1996). This vegetation pattern is (II) The combined effect of additional environmental generally controlled by low precipitation (< 300 mm), factors (differing from the topo-climatic factors high evapotranspiration and relief-controlled differences controlling forest distribution) control the actual in insolation in the mid-latitudes (Schlütz et al. 2008; living-tree biomass in the study area. Hais et al. 2016). Riverine forests consist of willow (III)Frequent forest fires, logging, and wood pasture (Salix), poplar (Populus), and larch (L. sibirica). Since strongly reduced the forest area and living-tree bio- these alluvial forests are supported by groundwater, they mass since prehistorical time. Thus, forests only are rather independent from local precipitation. Pleisto- partially cover the potential forest area that can be cene dune fields with scattered individual old larch trees deduced from climatic and topographic conditions. are abundant in the basins. Many local forest and steppe fires occur during summer (Goldammer 2007; Hessl Materials and methods et al. 2012). Severe forest fires in 1996 and 2002 Study area destroyed extensive forests. Many of these former forest The study area is located on the northern edge of the areas have not yet regrown. A timber factory and forest Khangai Mountains near the town Tosontsengel in tracks were established in the Tosontsengel region dur- northern central Mongolia (98°16′ E, 48°46′ N) (Fig. 1). ing Soviet times to facilitate intensified forest Klinge et al. Forest Ecosystems (2021) 8:55 Page 4 of 20 Fig. 1 Study area. a) Overview of Mongolia with position of the map shown in b) (black rectangle). b) Location of the study area in the forest- steppe of northern central Mongolia. Forest distribution was adapted from Klinge et al. (2018), burnt forest area (2000–2018) was adapted from Hansen et al. (2013). The digital elevation model (DEM) was created from SRTM (Shuttle Radar Topography Mission) data. The black rectangle in b) indicates the position of the image shown in c). c) True-colour satellite image of the study area near Tosontsengel (Landsat 8, September 22, 2014) Klinge et al. Forest Ecosystems (2021) 8:55 Page 5 of 20 Fig. 2 Climate of the study area around the town Tosontsengel (black circle). Data from the CHELSA V1.2 dataset, measuring period 1979–2013 (Karger et al. 2018), the shaded relief illustration is based on TanDEM-X data exploitation since the 1960s. Former clear-cutting is still 23, 1986 to delineate the distribution of forest prior to documented by rotted tree stumps inside the forests. In- extensive forest destruction through fires. This image dustrial logging was abandoned after the political change from 1986 was the best image for the period before the in the early 1990s, but has been resumed to some extent. onset of extensive forest fires. We determined the actual Illegal logging happens selectively inside of forests and forest area by integrating several scenes of Landsat 8 affects individual remnants in burnt areas, but no exten- (May 14, 2013; June 20, 2015) and Sentinel 2 (Sep 14, sive clear-cutting occurs. In addition, the local popula- 2016; Sep 19, 2017). After the intersection of the classifi- tion extracts selectively fuelwood from the forests. cation, we went through an intensive visual check of all forest polygons to delete and edit wrong classified areas. Remote sensing analysis of forest distribution, forest In the forest-steppe, distinct boundaries between forest categories and landscape units patches and grassland allowed for highly accurate forest This work step was crucial, as the main aim of this study maps. As shown by Klinge et al. (2015), image classifica- was to identify a possible mismatch between actual and tion combined with manual post-editing leads to an potential forest area. Input data for this work step in- overall accuracy of > 0.99. We used the difference in for- cluded Landsat 5, Landsat 8 and Sentinel 2 images est area between the images of 1986 and 2017 to work (Fig. 3, first line). For the mapping of forest areas, we ap- out the burnt forest area for this period. Because of the plied a semi-automatic approach after Klinge et al. different spatial resolutions of the satellite images, diver- (2015). The manual mapping of forest stands from satel- gences less than 20 m at the forest edges were neglected. lite imagery was supported by a previous supervised We distinguished several forest categories that occur maximum likelihood classification, where training sam- in different landscape units. Based on the proportion be- ples were distinguished for forest, steppe and water bod- tween forest and steppe, we differentiated between forest ies. We used a Landsat 5 satellite image from September stands in forest-dominated area and steppe-dominated Klinge et al. Forest Ecosystems (2021) 8:55 Page 6 of 20 Fig. 3 Workflow of this study area. The area, where the mountain tops reach above the slope, and insolation. We applied a GIS tool to estimate upper treeline, was defined as high-mountain area. For- the cumulative solar radiation input for the period Mai- ests in flat areas (< 2°) along rivers were interpreted as September 2017, which served as mean growing season alluvial forests. Forest stands on dunes were directly (MGS). These parameters were extracted for forest area identified in the satellite images. Furthermore, we distin- using the forest map of 1986, in order to identify rela- guished four forest-size classes in the forest-dominated tionships between potential forest distribution and relief. and steppe-dominated areas, respectively: F1, G1 ≤ 0.1 This approach allowed us to determine suitable topo- 2 2 2 2 km ; F2 = 0.1–1km ;F3=1–5km ;F4 ≥ 5km , using a graphic conditions and topographic thresholds for forest spatial buffer of 30 m to distinguish the forest edges growth from the respective value range. We produced a from the interior. These classification schemes were map of potential forest area based on relief parameters adapted from Dulamsuren et al. (2016, 2019). (PFA ), assuming that forest growth is possible in all areas within the topographic thresholds for forest Determination of suitable topographic conditions and growth, applying the approach of Klinge et al. (2015). topographic thresholds for forest growth Relief leads to variations of local climate, which is a par- Determination of suitable climatic conditions and climatic ticularly important aspect in semiarid central Asia thresholds for forest growth (Klinge et al. 2015). Therefore, we created a digital eleva- Similarly to the determination of PFA described above, tion model (DEM) based on TanDEM-X data (Fig. 3, we also analysed PFA based on relationships between first line). Due to high horizontal (10 m × 10 m) and ver- forest distribution and climatic parameters (PFA ). For tical (< 1 m) resolution of the DEM, the calculated ter- this purpose, we used climate data from the CHELSA rain surface was distorted by the forest canopy, V1.2 dataset (Karger et al. 2018), which we resampled especially at the edges of the forest stands. Therefore, we from the originally 30-arcsec resolution to obtain 30-m used the map of the actual forest area obtained from sat- resolution (Fig. 3, first box). This reanalysed climate ellite imagery to correct the DEM in forests. dataset enabled us to consider terrain parameters and The corrected DEM allowed us to extract various wind effect, and thus allowed us to obtain an improved topographic parameters, including elevation, aspect, representation of climate conditions in relief terrain Klinge et al. Forest Ecosystems (2021) 8:55 Page 7 of 20 (Karger et al. 2017). We calculated mean annual precipi- aboveground and belowground living-tree biomass. Dif- tation (MAP), mean annual potential evaporation, and ferences in the estimates of the two functions are dis- mean growing-season temperature (MGST, as the aver- cussed in Dulamsuren et al. (2016). We presumed that age of the monthly mean temperatures between May the increase of total tree biomass over the four-year and September) for the period 1973–2013. Temperature period of biomass-data collection in the field was less and precipitation were largely independent (Fig. 2), than the precision of the allometric method. Forest whereby temperature followed a vertical gradient, and stands where tree stumps indicated logging, were precipitation showed an additional longitudinal gradient mapped as forests with “logging”, whereas forest stands caused by the westerlies. Potential evaporation was so without tree stumps were mapped as forests with “no closely correlated with MGST that we did not include it logging”. Forest stands, where burnt bark and/or charred as an additional parameter for delineating PFA . wood indicated former fire events, were mapped as hav- The climatic parameters were extracted for forest area ing “fire indicators”, those without as having “no fire in- using the forest classification of 1986, in order to iden- dicators”. Ground vegetation structure, soil profiles and tify relationships between potential forest distribution detection of permafrost provided auxiliary data. Ground- and climate, and to determine suitable climatic condi- vegetation structure was important for assessing, which tions and climatic thresholds for forest growth. The portion of the NDVI of the forest sites was contributed intention was to produce a map of potential forest area by ground vegetation, because tree-canopy closure was based on climatic parameters (PFA ), assuming that for- less than 53%. In 24 of the plots, we measured leaf area est growth is possible in all areas within the climatic index (LAI) using a LI-COR Plant Canopy Analyzer thresholds for forest growth, thus, using the same ap- LAI-2200 C (Licor Biosciences, USA). Soil profiles were proach as for the PFA map. used to distinguish soils developed in sandy sediment and slope debris, and to detect permafrost, as permafrost Tree-biomass analysis is a crucial factor for forest distribution due to its During fieldwork in the years 2014–2018, we determined impounding effect for meltwater from seasonal ground living-tree biomass on 20 m × 20 m plots (Fig. 3, last box ice, keeping the meltwater available for trees. Permafrost in first line), by measuring tree diameter at breast height distribution was not used for biomass and PFA delinea- (dbh) and tree height of all living trees exceeding a tion, but its ecological feedback represents a relevant height of 4 m. In addition, we counted seedlings, saplings secondary parameter as shown by Klinge et al. (2021). and trees < 3 m, whereas we did not include the dead Classification of the plot data according to other influ- biomass. We used either a Vertex IV ultrasonic clinom- encing factors (Table 1) in addition to relief and climate, eter and T3 transponder (Haglöf, Långsele, Sweden) or a allowed for extracting effects of these factors on tree True Pulse 200 laser rangefinder (Laser Technology, biomass through statistical analysis. Inc., USA) for measuring tree height. Stem diameter was Dulamsuren et al. (2019) already analysed tree biomass calculated from stem circumference as measured with a in the interior of larch forests on slopes in the same measuring tape. Plot dimensions were determined by study area, thereby focusing on larch stands in the measuring tape, and plot-corner positions were mea- optimum stage of the forest development cycle (Jacob sured by GPS at an accuracy of ~ 3 m. For spatial refer- et al. 2013; Feldmann et al. 2018) and excluding distur- ence of the plot data, the centre between the four plot bances from fire or logging. Complementary to that corners was calculated. For statistical correlation analysis study, we also included larch stands influenced by vari- between biomass and remote sensing data, mean values ous factors, in order to also address the response of tree were interpolated of the pixels located within a 400-m biomass to these factors. In doing so, we also tested the circle around this point (Fig. 3, upper left side). In this potential of remote sensing techniques for upscaling way, we analysed 140 plots, including forest- and steppe- plot-based data to the landscape level. We used tree- dominated landscapes, forest edges and interiors, toe biomass data of 30 L. sibirica plots on slopes from slopes, mid-slopes, upper slopes, pristine and exploited Dulamsuren et al. (2019) and added tree-biomass data forests, as well as forest stands on slopes of different as- from forest edges and further L. sibirica plots of differ- pects and of different forest-stand sizes. We selected the ent forest-stand sizes (classes F1/G1 to F4) and forest- biomass plots according to their representativeness for a to-grassland ratios (forest-dominated area with classes larger surrounding area, to minimise discrepancies be- F1 to F4 vs. steppe-dominated area with class G1). tween field data and remote sensing data. In addition to these forests on slopes, we also analysed We applied the two allometric functions for Siberian larch forests on alluvial sand in floodplains. However, larch (L. sibirica) in Mongolia published by Battulga their limited size did not allow for obtaining separate et al. (2013) and Dulamsuren et al. (2016), and used the datasets for forest interior and edge. Altogether, we dis- mean of the results of both equations to estimate the tinguished 12 larch-stand categories, including the Klinge et al. Forest Ecosystems (2021) 8:55 Page 8 of 20 Table. 1 Mean living-tree biomass (above and belowground) for different forest categories affected by various factors. Plots, where a possible influence of a certain factor could not be assured, were excluded from the respective part of the analysis. SE = standard error, n = number of plots F1 SE n F2 SE n F3 SE n F4 SE n G1 SE n Total 198.7 11.2 29 208.0 11.8 31 212.5 13.9 31 182.0 12.9 34 142.2 10.7 10 Forest interior 219.3 14.0 18 218.2 12.6 26 220.6 13.4 26 181.7 13.8 30 145.4 15.0 7 Forest edge 165.1 13.7 11 155.2 21.0 5 170.7 46.2 5 184.3 35.2 4 134.8 2.7 3 Difference 54.2 63.0 49.9 −2.6 10.6 No fire indicators; 211.5 22.3 2 210.7 18.1 5 219.0 12.8 4 170.7 14.3 15 no logging No fire indicators 204.8 18.7 14 206.2 12.8 14 194.3 17.4 21 168.3 13.3 23 122.1 2.7 2 Fire indicators 179.5 15.2 9 171.6 20.9 11 231.3 14.4 6 175.5 39.7 5 154.9 1 Difference 25.2 34.6 −36.9 −7.2 No logging 211.5 22.3 2 234.3 11.9 8 238.8 23.8 6 180.4 13.7 18 134.1 4.0 2 Logging 197.8 11.9 27 198.9 14.4 23 207.4 16.6 24 183.8 22.7 16 144.3 13.2 8 Difference 13.7 35.4 31.4 −3.4 Slope debris 211.0 18.4 13 207.8 16.5 18 196.3 18.5 19 176.7 14.3 28 Sand layer 188.8 13.3 16 204.5 15.8 14 238.3 18.3 12 214.4 24.5 7 Difference 22.2 3.4 −42.0 −37.7 additional influencing factors forest interior / forest remnants. In the forest-dominated central and north- edges of the slope-forest categories F1, F2, F3, F4 and eastern parts of the study area, the most extensive burnt G1, and the floodplain forests as independent variables, forest areas were in the upper mountains. Only few for- and differentiating between logged / not logged and est stands burnt down in the steppe-dominated areas in burnt / not burnt forest stands as covariates, based on the north-western and eastern parts of the study area. the presence / absence of tree stumps and fire scars. Usual windthrow that creates single deadwood inside Permafrost distribution natural forests was not considered as disturbance. Permafrost was restricted to large forest stands on slopes We calculated the mean living-tree biomass for each in the forest-dominated area and high-mountain area, as forest category (as affected by the diverse factors), con- observed in our soil profiles (Klinge et al. 2021). Under sidering, e.g., logging, fire indicators, and topsoil condi- large forest stands (forest-size class F4) on north-facing tions (Table 1, Table S1). We checked the tree-biomass slopes, the permafrost was rich in ice and occurred data of each forest category for normal distribution and already at shallow depth, whereas east- and west-facing tested the differences in living-tree biomass between the slopes had only small patches of permafrost that started forest categories for statistical significance using Dun- at depths of more than 1 m. There was no field evidence can’s multiple range test calculated in SPSS. for permafrost under fragmented forest stands (G1, F1, We multiplied the area of each forest category with F2) and burnt forests (Klinge et al. 2021). the mean tree biomass of that forest category, using three scenarios, namely i) the actual forest area, ii) the PFA delineation based on relief parameters (PFA ) forest area of 1986, and iii) the potential forest area The upper treeline in the study area rises from 2400 m (PFA). Delineation of potential alluvial-forest area was in the north to 2600 m a.s.l. in the south (Klinge et al. not feasible, because the alluvial-forest distribution pat- 2018). Since 1986, forest fires in the upper mountains tern was largely controlled by the erosion-deposition dy- led to a decline in the mean elevation range of forests namics of the braided rivers. (95%) to 1600–2400 m a.s.l.. South-facing slopes in the forest-steppe are generally covered by steppe vegetation; Results they may be partially forested above 2100 m a.s.l. (Fig. 5). Spatial patterns of forest-fire and permafrost distribution Forests also occur in areas of maximum MGS insolation, Forest-fire distribution which demonstrates that insolation is no limiting factor The high-mountain area in the southern part of the for forest growth in the study area. study area lost the largest portion of forest through fire We delineated the PFA by clipping the area, where all over the past decades (Fig. 4). Its formerly large forests three parameters ‘aspect’ (no forest below 2100 m a.s.l. turned into numerous small and fragmented forest on slopes with aspect 135°–225°), ‘slope gradient’ (0– Klinge et al. Forest Ecosystems (2021) 8:55 Page 9 of 20 Fig. 4 Landscape units in the study area, with actual forest area and burnt forest area (reference year 2018). The shaded relief illustration is based on TanDEM-X data 25°), and ‘elevation’ (upper treeline: 2400 m a.s.l. in the presently forested area on steep slopes, from the valleys north, 2600 m a.s.l. in the south) allowed for tree growth up to the upper treeline at 2600 m a.s.l.. (Fig. 3). The most striking outcome of this PFA projec- tion was a much more extensive forest cover on toe PFA delineation based on climatic parameters (PFA ) slopes and pediments, which are at present generally The spatial resampling of the climate data by linear covered by steppe vegetation (Fig. 6). In the high- interpolation produced some noise, because small topo- mountain area, the estimated PFA exceeded the graphic variations could not be considered. Therefore, r Klinge et al. Forest Ecosystems (2021) 8:55 Page 10 of 20 Fig. 5 Top - Change of maximum MGS insolation (left axis) with elevation (MGS = mean growing season, May–September). Solid lines = MGS insolation on forest area, dashed lines = MGS insolation on the total land surface. Bottom - Forest area in hectares (right axis) in 1986 plotted against elevation. * The forest area on south-exposed slopes is shown in ha × 10 we did not use the obtained climate dataset to deduce except for a difference of the G1 forest edge plots to the the climatic thresholds for forest growth by histogram F1 and F2 interior plots (Fig. 8). Forests of the size clas- − 1 analysis. Instead, we derived these thresholds from step- ses F1, F2, and F3 had 50–63 Mg·ha less living-tree wise adjustment of the climatic thresholds until the biomass at forest edges than in their interiors; only the spatial pattern of temperature and precipitation with forest-size class F4 showed no distinct difference in tree positive conditions for tree growth included all forest biomass between forest edges and interior (Table 1). stands of 1986. The climatic thresholds obtained from Small fragmented forests G1 in the steppe-dominated − 1 this approach are listed in Table 2. The obtained PFA area had up to 70 Mg·ha less tree biomass than those showed larger forest areas on the upper south-facing in the forest-dominated area. Logging, forest fire and slopes and small flat summits. It matched well with the sediment type did not significantly influence living-tree upper treeline in the high-mountain area (Fig. 7). Com- biomass. The large proportion of forest with logging in pared to the PFA , the PFA did not extend as far down the categories of small fragmented forests (size classes r c into the basins, which may be due to low precipitation F1 and G1) pointed to a higher exploitation pressure on there. these small forests compared to larger stand sizes. Plot-based tree-biomass data Remote-sensing analysis based on tree-biomass estimates Mean tree height ranged between 12 and 20 m, whereas We tested several NDVI datasets and various topo- the maximum heights of single trees reached up to 32.7 graphic parameters for significant single or multi- 2 − 1 m. Stand basal area ranged from 5 to 91 m ·ha , with correlations with measured tree biomass, as such corre- 2 − 1 an average of 38.8 m ·ha . Mean tree ages were 100– lations would allow for interpolating tree biomass to 200 years, whereby maximum tree ages reached up to landscape scale. However, we found no statistically sig- 380–413 years. Living-tree biomass in larch forests on nificant correlations (r > 0.5, p < 0.05). One problem for − 1 slopes ranged between 25 and 380 Mg·ha . Maximum NDVI analysis was the low number of multispectral sat- − 1 tree biomasses of 440–688 Mg·ha were found in for- ellite images with < 10% cloud cover taken during sum- ests on floodplains, whereas larch trees on sand dunes mer. Another problem was the weak correlation only formed open woodlands with less tree biomass (48 between needle volume and tree biomass, as leaves and − 1 Mg·ha , n = 1). In all stand-size classes of the forests of needles provide the chlorophyll signal in multispectral the forest-dominated area, tree-biomass means and me- satellite images. Danilin and Tsogt (2014) stated that − 1 dians were within the range 180–220 Mg·ha ; max- needle biomass is independent of the average age of a − 1 imum tree biomass exceeded 320 Mg·ha (Table 1). larch stand, whereas tree biomass increases with tree Duncan’s multiple range test did not proof statistically age. The absence of a significant correlation between leaf significant differences between the forest categories, area index (LAI) and tree biomass measured on 24 plots Klinge et al. Forest Ecosystems (2021) 8:55 Page 11 of 20 Fig. 6 Forest distribution in 1986 and potential forest areas (PFA) delineated based on climate and relief. The shaded relief illustration is based on TanDEM-X data confirmed this statement (Fig. 9). Overall, the statistical Table. 2 Thresholds of mean growing-season temperature relationships between NDVI, needle volume and tree (MGST) and mean annual precipitation (MAP) used for PFA c biomass were poor. delineation Aspect Synthesis of the results of the forest-distribution and North East South West tree-biomass assessments The size of the study area was 6355 km . A closed third Maximum MGST (°C) 10.8 10.4 8.6 10.0 of the total area (1898 km ) was still forested in 1986. Minimum MGST (°C) 6.5 6.5 6.5 6.5 Since then, the forested area declined to 1086 km (ac- Maximum MAP (mm) 320 310 340 340 tual forest area). The delineated PFA yielded 3168 Minimum MAP (mm) 160 170 290 165 (PFA ) and 3553 km (PFA ), respectively. Details on the c r Klinge et al. Forest Ecosystems (2021) 8:55 Page 12 of 20 Fig. 7 Frequency-distribution curves of mean growing-season temperature (MGST) and mean annual precipitation (MAP) in the study area. Solid lines = forest area, dashed lines = total area. Climate data source: CHELSA V1.2 (Karger et al. 2017), period 1979–2013, spatially resampled to 30 m by linear interpolation differences between actual forest area, forest area in forest-dominated area, because both PFA projections 1986, and PFA, are listed in Table 3. The largest portion had resulted in forest-dominated landscapes (Table 5). of the actual forest area falls into the forest-size class F4 The actual tree biomass in the study area was 57% of the (8.6%) in the forest-dominated area. Prior to the large one in 1986, corresponding to 30% and 34% of the tree fire events, this forest-size class was also widespread in biomass estimated for the PFA and PFA , respectively r c the high-mountain area (9.2%). Altogether, a forest area (Table 5). The greatest losses of living-tree biomass due of 812.5 km (12.8% of the total study area and 42.8% of to forest fires since 1986 were detected in the large for- the formerly forested area) was destroyed by fire since ests (size class F4) of the forest-dominated area and the end of the last century. The burnt forest area was high-mountain area, whereas tree-biomass losses negligible in the steppe-dominated area and small in the through fire were less severe in the steppe-dominated forest-dominated area, but it amounted up to 95% of the area and alluvial forests. formerly forested area in the high-mountain area. The ratios between forest interior and forest edge (F /F ) for Discussion i e each forest-size class were relatively consistent for the Forest distribution different landscape units (Table 3). The main natural factors that control the spatial distri- Due to the low correlation between topographic pa- bution and vigour of forests in the Mongolian forest- rameters, NDVI and living-tree biomass (Fig. 9), estima- steppe are low precipitation and high evapotranspiration. tion of the total living-tree biomass in the study area by The latter depends on insolation, which in turn varies use of regression functions was not feasible. Therefore, with relief, resulting in a lack of forests on south-facing we estimated the total living-tree biomass by multiplying slopes (Dulamsuren and Hauck 2008; Hais et al. 2016; the specific mean tree biomass of each forest category by Klinge et al. 2018). Where forests occur, the forest can- the area of that forest category in 1986 and 2018 (Ta- opy fosters dense ground vegetation and an organic sur- bles 3, 4, 5). The total living-tree biomass of the PFA face layer that insulates the soil from warm air during was calculated based on the mean living-tree biomass of summer (Dashtseren et al. 2014). In this way, forests Klinge et al. Forest Ecosystems (2021) 8:55 Page 13 of 20 − 1 Fig. 8 Boxplots of tree biomass (Mg·ha ) of forests differing with respect to edge effects, influence of fire, logging and sediment type. Horizontal line = median, bars = quartiles, whiskers = range, dots = outliers. Means sharing a common letter, do not differ significantly (Duncan’s multiple range test, p ≤ 0.05) support discontinuous permafrost (Klinge et al. 2021). In Nevertheless, steppe vegetation predominates on the turn, permafrost helps trees to survive summer droughts, pediments, mainly because of herbivore grazing (Hilbig as it prevents meltwater that is released above the 1995). The climate-limited potential forest area (PFA ) permafrost table from percolating below the rooting obtained from this study yielded a lower treeline where zone (Sugimoto et al. 2002). Recognising these mutual dry conditions of the basins prevent tree growth, coin- relationships is crucial for understanding the patterns of ciding with the present lower forest boundaries. Exist- forest distribution and tree biomass in the Mongolian ence of forests below the threshold of 160 mm MAP can forest-steppe. However, this causal network alone cannot be explained by additional water supply through lateral explain the present forest-distribution pattern, as it is water fluxes, cumulating in concave positions and toe additionally influenced by other - mostly anthropogenic slopes (Klinge et al. 2021). The PFA moreover suggests - factors that may lead to a discrepancy between the ac- a potential for greater forest areas on south-facing tual and potential forest area (PFA). slopes. This mismatch reconfirms that MAP and MGST alone cannot explain forest distribution, which is a result Potential forest area (PFA) of a more complex causal network as explained in the The relief-limited potential forest area (PFA ) obtained beginning of this chapter. Short growing seasons and from this study suggests a potential for forest expansion, long-lasting snow cover prevent the expansion of forests both downslope towards the basins, and upslope towards into upper valleys and onto the mountain plateaus of the the high-mountain area. Pediments that widely cover the high-mountain area in the south. Another limiting factor toe slopes in the study area generally provide suitable there is the extensive use of alpine meadows as summer geoecological conditions for tree growth, as confirmed pastures. The upper treeline rises from 2400 m a.s.l. in by several existing small forest stands there. the north to 2600 m a.s.l. in the south of the study area. Klinge et al. Forest Ecosystems (2021) 8:55 Page 14 of 20 Fig. 9 Relationships between leaf area index (LAI) and living-tree biomass (left axis, blue dots), and between LAI and NDVI (right axis, red dots) of 24 plots. NDVI = mean NDVI obtained from seven Sentinel 2 images (Fig. 3) Thereby, small treeless areas on the flat summits of the the past 450 years based on tree ring analysis. In agree- northern mountains may result from the so-called “sum- ment with results from the Tuva region in southern Si- mit effect” (Körner 2012), i.e., particularly harsh condi- beria (Ivanova et al. 2010), the authors did not detect an tions near summits, rather than from a true upper increase in fire frequency during the last decades, but treeline. The projected PFAs suggest more large forests fires became more severe due to drier conditions. The and considerably less small, fragmented forest stands in limited contemporaneity of fire events at different sites the steppe-dominated area in the northern part of the pointed to fire-raising by humans (Hessl et al. 2012). On study area, assuming potential forest-dominated area the other hand, human impact may also lead to reduced there. A shift from potential forest-dominated area to destructiveness of fires, as wood gathering and intensive the presently observed steppe-dominated area may have grazing of livestock reduce available fuel for fires been partially triggered by natural factors such as fire, (Umbanhowar et al. 2009; Hessl et al. 2012). windbreak, insect calamities, and drought, but logging Although the most extensive forest fires in this area and forest pasture most likely caused major forest losses occurred already in 2002, forests have not yet re- in this area. Given the permafrost-promoting effect of established in many burnt forest areas. The most exten- large forests, the proposed forest-dominated area sce- sive burnt forest areas are located in the large forests of nario for the northern part of the study area would in- the high-mountain area and in the upper mountains in volve also greater abundance of permafrost in this area. the forest-dominated area. In contrast, only few burnt forest areas occur in the small and fragmented forests of Impact of fire the steppe-dominated area. We conclude that forest The forest area that burnt down between 1986 and 2017 fragmentation in the steppe-dominated area prevents amounts to 12.8% of the total study area. The loss of forest fires from passing over into neighbouring forest living-tree biomass since the last century adds up to stands and keeps fires rather isolated. The decrease of roughly 15 million tons, which represents more than large forests (size class F4) by fire led to an increase in 45% of the former tree biomass. Nyamjav et al. (2007) small, fragmented forest stands (size class F1), represent- stated that 95% of the actual forest destruction was ing remnants of the former large forests. This change in- caused by forest fires, whereas 5% was due to logging. duced loss of permafrost in these areas. Surviving larch The authors reported an increase of fire events in trees in the remaining forest remnants show enhanced Mongolia during the past decades. Goldammer (2002) fructification (Danilin and Tsogt 2014). Their important assumed that most of the fires were caused by human role as nuclei for forest regeneration is demonstrated by activities. Hessl et al. (2012) investigated fire history over numerous seedlings and saplings growing in the direct Klinge et al. Forest Ecosystems (2021) 8:55 Page 15 of 20 Table. 3 Dimensions (km ) and relative portions (%) of different forest categories and landscape units in the study area at present (actual forest area) and in 1986, prior to the large forest fires Actual forest area Burnt Forest distribution in 1986 Steppe forest Forest-size class 1 2 3 4 1 2 3 4 Absolute area of landscape type (km ) Steppe-dominated area Interior 4.74 18.25 21.51 7.29 3.88 5.59 19.65 22.90 7.92 1042.29 Edge 11.07 12.33 7.65 1.69 10.28 16.19 6.92 1.37 Ratio I/E 0.43 1.48 2.81 4.32 0.54 1.21 3.31 5.80 Forest-dominated area Interior 12.66 82.93 144.04 440.35 115.25 13.90 87.52 120.78 606.55 1907.62 Edge 43.55 52.56 44.21 107.34 32.10 42.10 32.12 105.36 Ratio I/E 0.29 1.58 3.26 4.10 0.43 2.08 3.76 5.76 High-mountain area Interior 1.73 7.51 6.40 693.33 3.74 29.70 65.66 491.08 1021.74 Edge 13.06 6.73 2.32 10.70 15.92 20.71 93.56 Ratio I/E 0.13 1.12 2.76 0.35 1.87 3.17 5.25 Alluvial forest 3.98 3.42 1.09 0.05 2.74 2.74 2.74 485.74 Dune 0.01 27.28 Relative proportion of landscape type (%) Steppe-dominated area Interior 0.07 0.29 0.34 0.11 0.06 0.09 0.31 0.36 0.12 16.40 Edge 0.17 0.19 0.12 0.03 0.16 0.25 0.11 0.02 Sum 0.25 0.48 0.46 0.14 0.25 0.56 0.47 0.15 Forest-dominated area Interior 0.20 1.30 2.27 6.93 1.81 0.22 1.38 1.90 9.54 30.02 Edge 0.69 0.83 0.70 1.69 0.51 0.66 0.51 1.66 Sum 0.88 2.13 2.96 8.62 0.72 2.04 2.41 11.20 High-mountain area Interior 0.03 0.12 0.10 10.91 0.06 0.47 1.03 7.73 16.08 Edge 0.21 0.11 0.04 0.17 0.25 0.33 1.47 Sum 0.23 0.22 0.14 0.23 0.72 1.36 9.20 Alluvial forest 0.13 0.001 0.04 0.07 0.02 7.64 Dune 0.43 surrounding of the forest remnants, in the shade of the Living-tree biomass old trees. Thus, a slow but steady re-immigration of In contrast to the close relationships of forest distribu- larch trees into the burnt area proceeds from these for- tion with relief and climate, living-tree biomass showed est remnants. It may take up to 200 years until a forest no significant correlation with topographic parameters. regenerates to its state prior to a fire (Nyamjav et al. It turned out that forests of the Mongolian forest-steppe 2007). have highly variable living-tree biomass. In addition to Fires occur frequently in semi-arid environment (Hessl natural impacts on forests (e.g., fire, windbreak, insect et al. 2012). Thus, L. sibirica is fire-adapted to a certain calamities, and drought), logging and forest pasture may degree. Its survival of a fire depends on the type of fire affect living-tree biomass. Alluvial forests exist where (crown, surface or ground fire), fire intensity, season, river channels hamper wood pasture, logging, and forest and soil moisture. The prevalent survival of forest stands fires. These alluvial forests usually consist of old larch in depressions, erosion channels, and on toe slopes dem- trees and have large tree biomass. Open larch forests on onstrates the importance of soil moisture for tree dunes are also made up by very old trees, but have low survival. stand density and tree biomass. Klinge et al. Forest Ecosystems (2021) 8:55 Page 16 of 20 Table. 4 Total living-tree biomass (10 g) in different forest categories of the study area Actual forest area Forest distribution in 1986 Forest-size class 1 2 3 4 Sum 1 2 3 4 Sum Steppe-dominated area Interior 68,892 398,195 474,487 132,398 1,073,972 81,348 428,835 505,209 143,948 1,159,341 Edge 149,217 191,236 130,499 31,053 502,005 138,668 251,237 118,171 25,152 533,229 Sum 218,109 589,431 604,986 163,450 1,575,976 220,016 680,073 623,380 169,100 1,692,569 Forest-dominated area Interior 277,732 1,809,560 3,177,481 8,000,534 13,265,306 304,883 1,909,714 2,664,408 11,020,218 15,899,224 Edge 718,971 815,498 754,423 1,977,866 4,266,758 530,041 653,229 548,092 1,941,327 3,672,689 Sum 996,703 2,625,058 3,931,904 9,978,400 17,532,064 834,925 2,562,943 3,212,500 12,961,545 19,571,913 High-mountain area Interior 37,963 163,881 141,084 0 342,928 82,118 648,002 1,448,389 8,922,284 11,100,793 Edge 215,657 104,498 39,519 0 359,674 176,719 247,038 353,449 1,723,920 2,501,126 Sum 253,620 268,379 180,603 0 702,602 258,837 895,041 1,801,838 10,646,204 13,601,919 Alluvial forest 350,149 353,095 Total sum 20,160,791 35,219,497 Comparison of tree biomass Sum 1986 1,313,778 4,138,056 5,637,718 23,776,849 34,866,401 Sum 2018 1,468,432 3,482,868 4,717,493 10,141,850 19,810,642 Loss by fire − 154,655 655,189 920,225 13,634,999 15,055,759 Percentage (%) −11.8 15.8 16.3 57.3 43.2 In attempts to assess biomass at landscape scale, the statistically significant correlation between NDVI and NDVI is commonly used as a biomass proxy. However, living-tree biomass. Instead, the NDVI proved to be a its suitability depends on the scale and data resolution. suitable indicator of the growing conditions for the en- Dulamsuren et al. (2016) successfully applied NDVI at tire forest vegetation (Erasmi et al. 2021). regional scale for biomass estimation in Mongolia. How- The differences in mean tree biomass between the forest ever, in our local-scale analysis we did not obtain a categories distinguished in our study were up to 85 Table. 5 Potential forest area (PFA) and living-tree biomass as controlled by climate (PFA ) and relief (PFA ) c r Potential forest area PFA Potential forest area PFA c r Forest-size class 1 2 3 4 Steppe 1 2 3 4 Steppe Area (km ) Interior 11.49 69.21 77.30 2590.66 2692.99 11.17 69.52 91.04 2940.99 2308.48 Edge 35.06 38.45 23.15 322.99 38.46 42.14 28.87 330.62 Ratio I/E 0.33 1.80 3.34 8.02 0.29 1.65 3.15 8.90 Tree biomass (×10 g) Total sum Total sum Interior 251,930 1,510,158 1,705,159 47,068,688 50,535,934 245,059 1,516,857 2,008,287 53,433,781 57,203,985 Edge 578,839 596,605 395,092 5,951,310 7,521,845 635,056 653,769 492,614 6,091,966 7,873,404 Sum 830,768 2,106,762 2100,250 53,019,998 58,057,779 880,115 2,170,625 2500,901 59,525,748 65,077,389 Comparison of tree biomasses (×10 g) PFA - 1986 − 483,009 −2,031,294 −3,537,468 29,243,149 23,191,377 −433,662 −1,967,431 −3,136,817 35,748,898 30,210,988 Percentage (%) −58.1 −96.4 −168.4 55.2 39.9 −49.3 −90.6 −125.4 60.1 46.4 PFA - 2018 −637,664 −1,376,105 −2,617,242 42,878,148 38,247,136 − 588,317 −1,312,242 −2,216,592 49,383,898 45,266,747 Percentage (%) −76.8 −65.3 −124.6 80.9 65.9 −66.8 −60.5 −88.6 83.0 69.6 Klinge et al. Forest Ecosystems (2021) 8:55 Page 17 of 20 − 1 Mg·ha . Our field measurements showed that the least be considered in the PFA projection. For example, large mean tree biomasses occurred in the forest-size class G1 forests (size class F4), as predominantly obtained from − 1 (142 Mg·ha ) of the steppe-dominated area and in the the PFA delineation, are more prone to severe fires than − 1 class F4 (182 Mg·ha ) of the forest-dominated area. The fragmented forest stands. In addition, due to the long- reduced living-tree biomass of the small, fragmented for- lasting human influence, reconstructing the natural pro- ests in the steppe-dominated area (G1) can be explained portion between steppe and forest in this region remains by enhanced forest use, reducing the stand basal area. As a major research challenge (Klinge and Sauer 2019). Hu- therefore solar irradiation on the ground is increased, man impact already started with the extinction of large there is no permafrost under these forests. The reduced herbivores like elephantine, and the reduction of wild living-tree biomass of the largest forests (size class F4) has animal herds since the Mesolithic period. It continued a different reason. There, the permafrost table approaches with the breeding of domestic animals and the develop- the surface and hinders deep rooting of trees. Trees are ment of pasture since the Neolithic period, which started therefore highly prone to windthrow, which explains the around 4.7 ka BP with the Afanasievo culture in the Altai reduced living-tree biomass in this forest category. Mountains (Kovalev and Erdenebaatar 2009). Furthermore, forest edges showed reduced living-tree Tchebakova et al. (2009) modelled potential vegetation biomass. Forest edges represent natural zones of forest changes across Siberia based on climate-change scenar- expansion and retreat (Sommer and Treter 1999), ios projecting warmer and drier climate. The authors whereby temporal climatic variations control these fluc- forecast an increase of forest-steppe and grassland areas. tuations. Forest edges may have a fringe of dead trees at They assume that drier conditions and larger amounts their outer boundary, and their outer boundary may also of fuel due to enhanced tree mortality will lead to an in- be dissected. In addition, logging and pasture is more in- crease in frequency and destructiveness of fires. tensive at the forest edges than in the interiors. Due to the lower tree density, the living-tree biomass is gener- Conclusions ally lower at the forest edges than in the interiors. How- A combination of tree-biomass determination, perma- ever, we found exceptions to this rule in the forest-size frost detection in soil profiles, remote sensing and classes G1 and F4, where the forest interiors and edges climate-data analysis allowed us to identify factors con- had similar tree biomasses. In the class G1, the tree bio- trolling larch-forest distribution and living-tree biomass masses of the small interior forest areas are similarly low in the northern Khangai Mountains, central Mongolia. as those of the forest edges. In the class F4, the forest The identified topographic and climatic thresholds for edges have large tree biomasses compared to all other forest growth enabled us to delineate the potential forest forest edges. This can be explained by the effect of area (PFA), which was much larger than the actual forest permafrost. A shallow permafrost table in the interior area. Forest fires destroyed 43% of the forest area and part of F4 forests causes reduced tree biomass in the for- 45% of the living-tree biomass in the study area over the est interior as described above. Towards the forest edges, period 1986–2017. They mostly affected large forest the depth of the permafrost table increases. There, dur- stands in the upper mountains. Permafrost, which was ing summer permafrost supplies meltwater to the trees widespread under large forests, disappeared soon after that are otherwise close to the climatic threshold of tree the destruction of a large forest stand. growth at the drier forest edges. This effect, together In contrast to forest distribution, living-tree biomass with higher precipitation in the upper mountains may showed no correlation with topographic and climatic pa- also explain the existence of forests on south-facing rameters. We found neither significant differences in slopes in the higher mountains (Fig. 5). living-tree biomass between forests with different fire Interestingly, forest stands that experienced non-lethal history, degree of exploitation, and soil properties, nor fire events or selective logging had similar tree bio- between most forest-size classes. Only forest edges and masses as pristine forests. A possible explanation is that small, fragmented forest stands had significantly less tree moderate thinning of forests may improve the growing biomass than all other forest categories. Neither non- conditions for the remaining trees, as it leads to reduced lethal fires nor selective logging seriously reduced living- competition for water and increased nutrient supply tree biomass. We conclude that these impacts remove from ash, and the melting permafrost leads to a tempor- tree biomass, but also stimulate growth of the remaining ary increase of soil moisture and allows for deeper trees by reducing competition. rooting. Based on relief thresholds for forest growth, we ob- 2 9 The estimated maximum tree biomass of the PFA tained a PFA of 3552 km with 65 × 10 g tree biomass, 9 2 (58–65 × 10 g) was twice the tree biomass in 1986 and based on climatic thresholds a PFA of 3113 km 9 9 (35 × 10 g) and three times the actual tree biomass of with 58 × 10 g tree biomass, corresponding to 323% 20 × 10 g. However, several relevant factors could not and 288% of the actual tree biomass, respectively. Klinge et al. Forest Ecosystems (2021) 8:55 Page 18 of 20 However, these estimates do not consider several rele- Declarations vant factors such as herbivore grazing and plant compe- Ethics approval and consent to participate tition. In addition, long-lasting human impact (at Not applicable. millennial timescale) plays an important role for the vegetation pattern as well, which needs further Consent for publication Not applicable. investigation. Competing interests Abbreviations The authors declare that they have no competing interests. a.s.l.: Above the sea level; dbh: Tree diameter at breast height; DEM: Digital elevation model; LAI: Leaf area index; MAP: Mean annual precipitation; Author details MGS: Mean growing season, may–september; MGST: Mean growing season Department of Physical Geography, Institute of Geography, University of temperature; NDVI: Normalized differentiated vegetation index; PFA: Potential Göttingen, Goldschmidtstraße 5, 37077 Göttingen, Germany. Applied forest area; SE: Standard error; SRTM: Shuttle radar topography mission Vegetation Ecology, Faculty of Environment and Natural Resources, University of Freiburg, Tennenbacher Str. 4, 79106 Freiburg, Germany. Institute of Farm Economics, Thünen Institute, Bundesallee 63, 38116 Supplementary Information 4 Braunschweig, Germany. Department of Biology, School of Arts and The online version contains supplementary material available at https://doi. Sciences, National University of Mongolia, Baga toiruu 47, Sukhbaatar duureg, org/10.1186/s40663-021-00333-9. Ulaanbaatar, Mongolia. Received: 20 April 2021 Accepted: 22 July 2021 Additional file 1: Table S1 Mean living-tree biomass (above and be- lowground) for different forest categories and site conditions. Plots, where site conditions were not clearly identified, were excluded from the respective part of the analysis. The two lower and higher forest-size clas- References ses in the forest-dominated area were combined for the statistical ana- Academy of Sciences of Mongolia, Academy of Sciences of USSR (1990) National lysis, because of the small dataset for forest edges. SE = standard error, Atlas of the Peoples Republic of Mongolia, Ulaanbaatar, Moscow n = number of plots. Underlined data are not representative because of Batima P, Natsagdorj L, Gombluudev P, Erdenetsetseg B (2005) Observed climate insufficient size of the respective dataset. Fig. S1 NDVI of forests in the change in Mongolia. AIACC Working Papers 12:1–25 study area. Arithmetic means of NDVI from seven Sentinel 2 satellite im- Battulga P, Tsogtbaatar J, Dulamsuren Ch, Hauck M (2013) Equations for ages (17.05.2018, 11.06.2018, 25.08.2018, 4.09.2018, 14.09.2018, 19.09.2017, estimating the above-ground biomass of Larix sibirica in the forest-steppe of 16.07.2016). The shaded relief illustration is based on TanDEM-X data. Mongolia. J Forest Res 24(3):431–437. https://doi.org/10.1007/s11676-013-03 75-4 Dagvadorj D, Natsagdorj L, Dorjpurev J, Namkhainyam B (2009) Mongolia Acknowledgements assessment report on climate change 2009, Ulaanbaatar, Mongolia We thank Ms. Daramragchaa Tuya from the Tarvagatai Nuruu National Park Danilin IM (1995) Structure and biomass of larch stands regenerating naturally (Tosontsengel Sum, Zavkhan Aimag, Mongolia) for her invaluable support of after clearcut logging. Water Air Soil Pollut 82:125–131. https://doi.org/10.1 our research. We wish to express our gratitude to our Mongolian colleagues 007/978-94-017-0942-2_14 Mr. Amarbayasgalan, Mr. Enkhjargal, Mr. Enkh-Agar, Ms. Munkhtuya. We Danilin IM, Tsogt Z (2014) Dynamics of structure and biological productivity of greatly appreciated their hospitality and help with the fieldwork. Our thanks post-fire larch forests in the northern Mongolia. Contemp Probl Ecol 7(2): also go to the German students Martine Koob, Tino Peplau, Janin Klaassen 158–169. https://doi.org/10.1134/S1995425514020036 and Tim Rollwage for their great support with the biomass measurements Dashtseren A, Ishikawa M, Iijima Y, Jambaljav Y (2014) Temperature regimes of during the fieldwork in Mongolia. the active layer and seasonally frozen ground under a forest-steppe mosaic, The German Aerospace Centre (DLR) liberally provided the TanDEM-X data Mongolia. Permafrost Periglac Proc 25(4):295–306. https://doi.org/10.1002/ for the study area (DEM_FOREST 1106). The fieldwork in 2014 was funded by ppp.1824 the Volkswagen Foundation in the frame of the project 87175 “Forest regen- DeLuca TH, Boisvenue C (2012) Boreal forest soil carbon: distribution, function eration and biodiversity at the forest-steppe border of the Altay and Khangay and modelling. Forestry 85(2):161–184. https://doi.org/10.1093/forestry/ Mountains under contrasting development of livestock numbers in cps003 Kazakhstan and Mongolia” granted to M. Hauck, Ch. Dulamsuren and C. Dulamsuren Ch, Hauck M (2008) Spatial and seasonal variation of climate on Leuschner. The subsequent work was funded by the Deutsche Forschungs- steppe slopes of the northern Mongolian mountain taiga. Grassl Sci 54(4): gemeinschaft (DFG), project number 385460422 approved to M. Klinge, D. 217–230. https://doi.org/10.1111/j.1744-697X.2008.00128.x Sauer and M. Frechen. Dulamsuren Ch, Hauck M, Bader M, Osokhjargal D, Oyungerel S, Nyambayar S, Runge M, Leuschner C (2009) Water relations and photosynthetic performance in Larix sibirica growing in the forest-steppe ecotone of Authors’ contributions northern Mongolia. Tree Physiol 29(1):99–110. https://doi.org/10.1093/ MK conceived the ideas; MK, ChD, FS, MH, UB and DS participated fieldwork treephys/tpn008 and collected the data; MK, ChD and SE analyzed the data; MK, ChD, MH and Dulamsuren Ch, Hauck M, Khishigjargal M, Leuschner HH, Leuschner C (2010a) DS wrote the paper. The author(s) read and approved the final manuscript. Diverging climate trends in Mongolian taiga forests influence growth and regeneration of Larix sibirica. Oecologia 163(4):1091–1102. https://doi.org/10.1 007/s00442-010-1689-y Funding Dulamsuren Ch, Hauck M, Leuschner C (2010c) Recent drought stress leads to This study was funded by the Volkswagen Foundation (project-no. 871759) growth reductions in Larix sibirica in the western Khentey, Mongolia. Glob and by the German Research Council (Deutsche Forschungsgemeinschaft, Change Biol 16:3024–3035. https://doi.org/10.1111/j.1365-2486.2009.02147.x DFG), (project no. 385460422). Dulamsuren Ch, Hauck M, Leuschner HH, Leuschner C (2010b) Gypsy moth- induced growth decline of Larix sibirica in a forest-steppe ecotone. Availability of data and materials Dendrochronologia 28(4):207–213. https://doi.org/10.1016/j.dendro.2009.05. The datasets used and/or analyzed during the current study are available 007 from the corresponding author on reasonable request. Dulamsuren Ch, Hauck M, Mühlenberg M (2008) Insect and small mammal Clime data are publicly available from the CHELSA data set (https://chelsa- herbivores limit tree establishment in northern Mongolian steppe. Plant Ecol climate.org/). 195(1):143–156. https://doi.org/10.1007/s11258-007-9311-z Klinge et al. Forest Ecosystems (2021) 8:55 Page 19 of 20 Dulamsuren Ch, Khishigjargal M, Leuschner C, Hauck M (2014) Response of tree- Khishigjargal M, Dulamsuren C, Lkhagvadorj D, Leuschner C, Hauck M (2013) ring width to climate warming and selective logging in larch forests of the Contrasting responses of seedling and sapling densities to livestock density Mongolian Altai. J Plant Ecol 7(1):24–38. https://doi.org/10.1093/jpe/rtt019 in the Mongolian forest-steppe. Plant Ecol 214(11):1391–1403. https://doi. Dulamsuren Ch, Klinge M, Bat-Enerel B, Ariunbaatar T, Tuya D (2019) Effects of org/10.1007/s11258-013-0259-x forest fragmentation on organic carbon pool densities in the Mongolian Klinge M, Böhner J, Erasmi S (2015) Modeling forest lines and forest distribution forest-steppe. Forest Ecol Manag 433:780–788. https://doi.org/10.1016/j. patterns with remote-sensing data in a mountainous region of semiarid foreco.2018.10.054 Central Asia. Biogeosciences 12(10):2893–2905. https://doi.org/10.5194/ bg-12-2893-2015 Dulamsuren Ch, Klinge M, Degener J, Khishigjargal M, Chenlemuge T, Bat-Enerel B, Yeruult Y, Saindovdon D, Ganbaatar K, Tsogtbaatar J, Leuschner C, Hauck Klinge M, Dulamsuren C, Erasmi S, Karger DN, Hauck M (2018) Climate effects on M (2016) Carbon pool densities and a first estimate of the total carbon pool vegetation vitality at the treeline of boreal forests of Mongolia. in the Mongolian forest-steppe. Glob Change Biol 22(2):830–844. https://doi. Biogeosciences 15(5):1319–1333. https://doi.org/10.5194/bg-15-1319-2018 org/10.1111/gcb.13127 Klinge M, Sauer D (2019) Spatial pattern of late glacial and Holocene climatic and Erasmi S, Klinge M, Dulamsuren Ch, Schneider F, Hauck M (2021) Modelling the environmental development in Western Mongolia - a critical review and productivity of Siberian larch forests from Landsat NDVI time series in synthesis. Quaternary Sci Rev 210:26–50. https://doi.org/10.1016/j.quascirev.2 fragmented forest stands of the Mongolian forest-steppe. Environ Monit 019.02.020 Assess 193(4):200. https://doi.org/10.1007/s10661-021-08996-1 Klinge M, Schneider F, Dulamsuren Ch, Arndt K, Bayarsaikhan U, Sauer D (2021) Feldmann E, Glatthorn J, Hauck M, Leuschner C (2018) A novel empirical Interrelations between relief, vegetation, disturbances, and permafrost in the approach for determining the extension of forest development stages in forest-steppe of Central Mongolia. Earth Surf Process Landforms 46(9):1766– temperate old-growth forests. Eur J Forest Res 137(3):321–335. https://doi. 1782. https://doi.org/10.1002/esp.5116 org/10.1007/s10342-018-1105-4 Körner C (2012) Alpine treelines: functional ecology of the global high elevation Goldammer G (2002) Fire Situation in Mongolia. International Forest Fire News tree limits. Springer, Basel, Heidelberg, New York, Dordrecht, London 26:75–83 Kovalev AA, Erdenebaatar D (2009) Discovery of new cultures of the bronze age Goldammer G (ed) (2007) International Forest Fire News. UNECE and FAO 36 in Mongolia according to the data obtained by the international central Goodale CL, Apps MJ, Birdsey RA, Field CB, Heath LS, Houghton RA, Jenkins JC, Asian archaeological expedition. In: Bemmann J, Parzinger H, Pohl E, Kohlmaier GH, Kurz W, Liu S, Nabuurs G-J, Nilsson S, Shvidenko AZ (2002) Tseveendorj D (eds) Current archaeological research in Mongolia, Bonn, pp. Forest carbon sinks in the northern hemisphere. Ecol Appl 12(3):891–899. 149–170 https://doi.org/10.1890/1051-0761(2002)012[0891:FCSITN]2.0.CO;2 Kowalkowski K, Starkel L (1984) Altitudinal belts of geomorphic pro¬cesses in the Hais M, Chytrý M, Horsák M (2016) Exposure-related forest-steppe: a diverse southern Changai Mts. (Mongolia). Stud Geomorphol Carpatho-Balcan 18:95– landscape type determined by topography and climate. J Arid Environ 135: 115 75–84. https://doi.org/10.1016/j.jaridenv.2016.08.011 Liu H, Park Williams A, Allen CD, Guo D, Wu X, Anenkhonov OA, Liang E, Sandanov DV, Yin Y, Qi Z, Badmaeva NK (2013) Rapid warming accelerates Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, Tyukavina A, Thau tree growth decline in semi-arid forests of inner Asia. Glob Change Biol 19(8): D, Stehman SV, Goetz SJ, Loveland TR, Kommareddy A, Egorov A, Chini L, st 2500–2510. https://doi.org/10.1111/gcb.12217 Justice CO, Townshend JRG (2013) High-resolution global maps of 21 - century forest cover change. Science 342(6160):850–853. https://doi.org/1 Lkhagvadorj D, Hauck M, Dulamsuren Ch, Tsogtbaatar J (2013a) Pastoral 0.1126/science.1244693 nomadism in the forest-steppe of the Mongolian Altai under a changing Hauck M, Dulamsuren Ch, Bayartogtokh B, Ulykpan K, Burkitbaeva UD, economy and a warming climate. J Arid Environ 88:82–89. https://doi.org/1 Otgonjargal E, Titov SV, Enkhbayar T, Sundetpaev AK, Beket U, Leuschner C 0.1016/j.jaridenv.2012.07.019 (2014) Relationships between the diversity patterns of vascular plants, lichens Lkhagvadorj D, Hauck M, Dulamsuren Ch, Tsogtbaatar J (2013b) Twenty years and invertebrates in the central Asian forest-steppe ecotone. Biodivers after decollectivization: Mobile livestock husbandry and its ecological impact Conserv 23(5):1105–1117. https://doi.org/10.1007/s10531-014-0648-z in the Mongolian forest-steppe. Hum Ecol 41(5):725–735. https://doi.org/10.1 Hessl AE, Ariya U, Brown P, Byambasuren O, Green T, Jacoby G, Sutherland EK, 007/s10745-013-9599-3 Nachin B, Maxwell RS, Pederson N, de Grandpré L, Saladyga T, Tardif JC Luyssaert S, Inglima I, Jung M, Richardson AD, Reichstein M, Papale D, Piao S, (2012) Reconstructing fire history in Central Mongolia from tree-rings. Int J Schulze E-D, Wingate L, Matteucci G, Aragao L, Aubinet M, Beer C, Bernhofer Wildland Fire 21(1):86. https://doi.org/10.1071/WF10108 C, Black K, Bonal D, Bonnefond J, Chambers J, Ciais P, Cook B, Davis K, Dolman A, Gielen B, Goulden M, Grace J, Granier A, Grelle A, Griffis T, Hilbig W (1995) The vegetation of Mongolia. SPB Acad Publ, Amsterdam Grünwald T, Guidolotti G, Hanson A, Harding R, Hollinger D, Hutyra L, Kolari IPCC (2013) Climate change 2013 - the physical science basis. Cambridge P, Kruijt B, Kutsch W, Lagergren F, Laurila T, Law B, Le Maire G, Lindroth A, University Press, Cambridge Loustau D, Malhi Y, Mateus J, Migliavacca M, Misson L, Montagnani L, Ivanova GA, Ivanov VA, Kukavskaya EA, Soja AJ (2010) The frequency of forest Moncrieff J, Moors E, Munger W, Nikinmaa E, Ollinger S, Pita G, Rebmann C, fires in scots pine stands of Tuva, Russia. Environ Res Lett 5(1):15002. https:// Roupsard O, Saigusa N, Sanz M, Seufert G, Sierra C, Smith M, Tang J, Valentini doi.org/10.1088/1748-9326/5/1/015002 R, Vesala T, Janssens I (2007) CO balance of boreal, temperate, and tropical Jacob M, Bade C, Calvete H, Dittrich S, Leuschner C, Hauck M (2013) Significance 2 forests derived from a global database. Glob Change Biol 13(12):2509–2537. of over-mature and decaying trees for carbon stocks in a central European https://doi.org/10.1111/j.1365-2486.2007.01439.x natural spruce forest. Ecosystems 16(2):336–346. https://doi.org/10.1007/s1 0021-012-9617-0 Miehe G, Miehe S, Böhner J, Kaiser K, Hensen I, Madsen D, Liu J, Opgenoorth L (2014) How old is the human footprint in the world’s largest alpine Jarvis P, Saugier B, Schulze E-D (2001) Productivity of boreal forests. In: Roy J, ecosystem? A review of multiproxy records from the Tibetan plateau from Saugier B, Mooney H (eds) Terrestrial global productivity. Academic Press, the ecologists’ viewpoint. Quaternary Sci Rev 86:190–209. https://doi.org/10.1 San Diego, pp 211–244. https://doi.org/10.1016/B978-012505290-0/50011-9 016/j.quascirev.2013.12.004 Karger DN, Conrad O, Böhner J, Kawohl T, Kreft H, Soria-Auza RW, Zimmermann Miehe G, Miehe S, Will M, Opgenoorth L, Duo L, Dorgeh T, Liu J (2007) An NE, Linder HP, Kessler M (2017) Climatologies at high resolution for the inventory of forest relicts in the pastures of southern Tibet (Xizang a.R., earth’s land surface areas. Sci Data 4(1):170122. https://doi.org/10.1038/sda China). Plant Ecol 194(2):157–177. https://doi.org/10.1007/s11258-007-9282-0 ta.2017.122 Mukhortova L, Schepaschenko D, Shvidenko A, McCallum I, Kraxner F (2015) Soil Khansaritoreh E, Dulamsuren Ch, Klinge M, Ariunbaatar T, Bat-Enerel B, Batsaikhan contribution to carbon budget of Russian forests. Agric For Meteorol 200:97– G, Ganbaatar K, Saindovdon D, Yeruult Y, Tsogtbaatar J, Tuya D, Leuschner C, 108. https://doi.org/10.1016/j.agrformet.2014.09.017 Hauck M (2017b) Higher climate warming sensitivity of Siberian larch in small Nyamjav B, Goldammer G, Uibrig H (2007) Fire situation in Mongolia. than large forest islands in the fragmented Mongolian forest steppe. Glob International Forest Fire News 36:46–66 Change Biol 23(9):3675–3689. https://doi.org/10.1111/gcb.13750 Pan Y, Birdsey RA, Fang J, Houghton R, Kauppi PE, Kurz WA, Phillips OL, Khansaritoreh E, Eldarov M, Ganbaatar K, Saindovdon D, Leuschner C, Hauck M, Shvidenko A, Lewis SL, Canadell JG, Ciais P, Jackson RB, Pacala SW, McGuire Dulamsuren C (2017a) Age structure and trends in annual stem increment of AD, Piao S, Rautiainen A, Sitch S, Hayes D (2011) A large and persistent Larix sibirica in two neighboring Mongolian forest–steppe regions differing carbon sink in the world’s forests. Science 333(6045):988–993. https://doi. in land use history. Trees 31(6):1973–1986. https://doi.org/10.1007/s00468-01 org/10.1126/science.1201609 7-1601-z Klinge et al. Forest Ecosystems (2021) 8:55 Page 20 of 20 Poulter B, Pederson N, Liu H, Zhu Z, D’Arrigo R, Ciais P, Davi N, Frank D, Leland C, Myneni R, Piao S, Wang T (2013) Recent trends in inner Asian forest dynamics to temperature and precipitation indicate high sensitivity to climate change. Agric For Meteorol 178-179:31–45. https://doi.org/10.1016/j.a grformet.2012.12.006 Richter H, Haase G, Barthel H (1963) Die Golezterrassen. Petermanns Geogr Mitt 107:183–192 Sakamoto K, Tomonari M, Ariya U, Nakagiri E, Matsumoto TK, Akaji Y, Otoda T, Hirobe M, Nachin B (2021) Effects of large-scale forest fire followed by illegal logging on the regeneration of boreal forests in Mongolia. Landscape Ecol Eng 17(3):267–279. https://doi.org/10.1007/s11355-021-00457-8 Schlütz F, Dulamsuren Ch, Wieckowska M, Mühlenberg M, Hauck M (2008) Late Holocene vegetation history suggests natural origin of steppes in the northern Mongolian mountain taiga. Palaeogeogr Palaeoclimatol Palaeoecol 261(3-4):203–217. https://doi.org/10.1016/j.palaeo.2007.12.012 Shvidenko AZ, Schepaschenko D (2014) Carbon balance of Russian forests. Siber Forest J 1:69–92 Sommer M, Treter U (1999) Die Lärchenwälder der Gebirgswaldsteppe in den Randgebieten des Uvs Nuur- Beckens. Erde 130:173–188 Sugimoto A, Yanagisawa N, Naito D, Fujita N, Maximov TC (2002) Importance of permafrost as a source of water for plants in east Siberian taiga. Ecol Res 17(4):493–503. https://doi.org/10.1046/j.1440-1703.2002.00506.x Tchebakova NM, Parfenova E, Soja AJ (2009) The effects of climate, permafrost and fire on vegetation change in Siberia in a changing climate. Environ Res Lett 4(4):45013. https://doi.org/10.1088/1748-9326/4/4/045013 Thurner M, Beer C, Santoro M, Carvalhais N, Wutzler T, Schepaschenko D, Shvidenko A, Kompter E, Ahrens B, Levick SR, Schmullius C (2014) Carbon stock and density of northern boreal and temperate forests. Glob Ecol Biogeogr 23(3):297–310. https://doi.org/10.1111/geb.12125 Treter U (1996) Gebirgs-Waldsteppe in der Mongolei. Geogr Rundsch 48:655–661 Tsogtbaatar J (2004) Deforestation and reforestation needs in Mongolia. Forest Ecol Manag 201(1):57–63. https://doi.org/10.1016/j.foreco.2004.06.011 Umbanhowar CE, Shinneman AL, Tserenkhand G, Hammon ER, Lor P, Nail K (2009) Regional fire history based on charcoal analysis of sediments from nine lakes in western Mongolia. Holocene 19(4):611–624. https://doi.org/1 0.1177/0959683609104039 Unkelbach J, Dulamsuren C, Punsalpaamuu G, Saindovdon D, Behling H (2017) Late Holocene vegetation, climate, human and fire history of the forest- steppe-ecosystem inferred from core G2-a in the ‘Altai Tavan Bogd’ conservation area in Mongolia. Veget Hist Archaeobot 27(5):665–677. https:// doi.org/10.1007/s00334-017-0664-5 Unkelbach J, Kashima K, Enters D, Dulamsuren C, Punsalpaamuu G, Behling H (2019) Late Holocene (Meghalayan) palaeoenvironmental evolution inferred from multi-proxy-studies of lacustrine sediments from the Dayan Nuur region of Mongolia. Palaeogeogr Palaeoclimatol Palaeoecol 530:1–14. https:// doi.org/10.1016/j.palaeo.2019.05.021 Yang Y, Wang Z, Li J, Gang C, Zhang Y, Zhang Y, Odeh I, Qi J (2016) Comparative assessment of grassland degradation dynamics in response to climate variation and human activities in China, Mongolia, Pakistan and Uzbekistan from 2000 to 2013. J Arid Environ 135:164–172. https://doi.org/10.1016/j.ja ridenv.2016.09.004

Journal

"Forest Ecosystems" – Springer Journals

Published: Aug 10, 2021

Keywords: Biomass; Fire; Forest-steppe; Geoecological factors; Mongolia; Permafrost

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

Geoecological parameters indicate discrepancies between potential and actual forest area in the forest-steppe of Central Mongolia

References (40)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies