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

Learn More →

Assessment of MODIS Surface Temperature Products of Greenland Ice Sheet Using In-Situ Measurements

Assessment of MODIS Surface Temperature Products of Greenland Ice Sheet Using In-Situ Measurements land Article Assessment of MODIS Surface Temperature Products of Greenland Ice Sheet Using In-Situ Measurements 1 , † 1 , † 2 1 1 1 Xiaoge Yu , Tingting Wang , Minghu Ding , Yetang Wang , Weijun Sun , Qinglin Zhang 1 , and Baojuan Huai * College of Geography and Environment, Shandong Normal University, Jinan 250061, China; 201914010518@stu.sdnu.edu.cn (X.Y.); 201914010317@stu.sdnu.edu.cn (T.W.); 615002@sdnu.edu.cn (Y.W.); 612033@sdnu.edu.cn (W.S.); zhangqinglin@stu.sdnu.edu.cn (Q.Z.) State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China; dingmh@cma.gov.cn * Correspondence: 616070@sdnu.edu.cn or huaibaojuan@126.com † These authors contributed equally to this work. Abstract: Satellite-based data have promoted the research progress in polar regions under global climate change, meanwhile the uncertainties and limitations of satellite-derived surface temperatures are widely discussed over Greenland. This study validated the accuracy of ice surface temperature (IST) from the moderate-resolution imaging spectroradiometer (MODIS) over the Greenland ice sheet (GrIS). Daily MODIS IST was validated against the observational surface temperature from 24 automatic weather stations (AWSs) using the mean bias (MB), the root mean square (RMSE), and the correlation coefficient (R). The temporal and spatial variability over the GrIS spanning from March 2000 to December 2019 and the IST melt threshold (1 C) were analyzed. Generally, the MODIS IST was underestimated by an average of 2.68 C compared to AWSs, with cold bias mainly occurring in winter. Spatially, the R and RMSE performed the better accuracy of MODIS IST on the northwest, northeast, and central part of the GrIS. Furthermore, the mean IST is mainly concentrated between 20 C and 10 C in summer while between 50 C and 30 C in winter. Citation: Yu, X.; Wang, T.; Ding, M.; Wang, Y.; Sun, W.; Zhang, Q.; Huai, B. The largest positive IST anomalies (exceeds 3 C) occurred in southwestern GrIS during 2010. IST Assessment of MODIS Surface shows the positive trends mainly in spring and summer and negative in autumn and winter. Temperature Products of Greenland Ice Sheet Using In-Situ Measurements. Keywords: MODIS; ice surface temperature (IST); validation; Greenland ice sheet (GrIS); variability Land 2022, 11, 593. https://doi.org/ 10.3390/land11050593 Academic Editor: Nir Krakauer 1. Introduction Received: 16 March 2022 6 2 The Greenland ice sheet (GrIS) covers an area of 1.7  10 km and is the second Accepted: 16 April 2022 largest ice sheet in the world after the Antarctic ice sheet [1]. The GrIS had been rela- Published: 19 April 2022 tively stable during the 1970s, 1980s, and the early 1990s with the accumulation and loss Publisher’s Note: MDPI stays neutral staying relatively balanced [2,3]. In recent decades, there has been much attention paid with regard to jurisdictional claims in to the GrIS due to its enhanced melting [4–6], particularly relevant to abnormal climatic published maps and institutional affil- conditions [7–9] and accelerated mass loss [10–13]. The enhanced ablation was measured iations. using infrared and passive-microwave records [14–18]. The analysis of in-situ observations, satellite data, and regional climate models (RCMs) all revealed the accelerating GrIS mass loss [10,19]. The GrIS mass balance has decreased more than twice since the 21st century comparing with the last century [20,21]. The mass loss of GrIS over the past decade is about Copyright: © 2022 by the authors. six times that of the 1980s and the GrIS has contributed 13.7  1.1 mm to the global sea level Licensee MDPI, Basel, Switzerland. rise since 1972 [2]. If the acceleration continues, several studies using numerical model This article is an open access article strategies to predict the future evolution of GrIS suggested that GrIS could contribute to a distributed under the terms and rise in global sea level of 2.3 cm by 2050 [22], and 3.5 cm to 76.4 cm by 2300 [23,24]. Surface conditions of the Creative Commons temperature is an important factor in controlling mass balance [25]. Recently, the increase Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ in GrIS mass loss is mainly because of the increased surface meltwater runoff [26], and pro- 4.0/). longed surface temperature elevate exacerbates melting [27]. Surface temperature affects Land 2022, 11, 593. https://doi.org/10.3390/land11050593 https://www.mdpi.com/journal/land Land 2022, 11, 593 2 of 14 the basal melt [28] as its rise is an important step in the melt process [29]. Thus, tracking IST can better account for surface melt, mass balance as well as ice sheet surface processes. Currently, observational stations, regional climate models, and remote sensing are the three main sources to obtain IST for studying [30–33]. The AWSs on the ice sheet provide observations of the GrIS, for example, the Greenland Climate Network (GC-Net) [1,34] and the Greenland Ice Sheet Monitoring Program (PROMICE), which monitors surface and air temperature [35–37]. Meanwhile, existing coastal and low-density weather stations on the ice sheet make it hard to get accurate surface temperature observations of the entire GrIS. The difficulties of manipulating station devices in severe polar regions also make it hard to obtain continuous and complete observations record. RCMs can compensate for data recorded at certain periods and locations which could not be obtained from weather stations, especially in the polar regions [38]. RCMs are reliable in the entire GrIS context, however, in the regional and local context, climate variability can be missed occasionally [39,40]. Hence, the way to measure the surface temperature through satellite remote sensing is more outstanding than AWSs and RCMs with high temporal and spatial resolution over the GrIS. Nonetheless, the most practical way to get a wide range of surface temperature data is through the satellite remote sensing [41]. Accurate determination of satellite-derived IST will improve modeling of ice sheet melt and other ice sheet processes and changes. The GrIS is sensitive to climate change and plays a vital role in the global climate [1]. Hall showed an increase IST of ~0.55  0.44 C/decade and the northeastern GrIS experi- enced the largest increase of ~0.95  0.44 C/decade [8]. Under these conditions, detailed and accurate monitoring of the surface temperature on the entire GrIS is of significance. Recently, a multilayer, daily IST-albedo-water vapor product of Greenland, spanning from March 2000 to December 2019, has been developed. This paper mainly validated the accuracy of the MODIS IST products based on 24 AWSs from PROMICE. A description of the data and methods used in this study were given in 2. In 3, the IST performance was first demonstrated, and then based on its favorable performance, the temporal and spatial variability of GrIS IST, IST anomalies, and trends were further reproduced. Finally, in 4, the MODIS IST melt threshold 1 C using daily ablation data from 8 AWSs was also analyzed. 2. Materials and Methods 2.1. Observational Data The 24 AWSs (Figure 1) data used in this study are from the PROMICE, which is operated by the Geological Survey of Denmark and Greenland (GEUS) in collaboration with the National Space Institute at the Technical University of Denmark [35]. The time series of the AWS has different observation periods and completeness (Table 1). In this study, the daily IST data was derived from measured downward and upward longwave irradiance and surface emissivity is set to 0.97 [42]. Mostly located in the coastal area of the GrIS, the spatial layout of the AWSs is uneven. Roughly 25% of the stations are located below 500 m and only 16.7% are above 1000 m (Table 1). Moreover, only EGP, KAN_U, and CEN are located at the accumulation area according to the ELA (equilibrium line altitude) of the corresponding basin in 2017 [43]. Table 1. Geographical characteristics of the meteorological stations used in this study. Station Latitude Longitude Elevation Elevation-ELA Start Date Name ( N) ( W) (m) (m) KPC_L * 79.91 24.08 370 707 17 July 2008 KPC_U 79.83 25.17 870 207 17 July 2008 EGP 75.62 35.97 2660 1583 1 May 2016 SCO_L 72.22 26.82 460 1192 21 July 2008 SCO_U 72.39 27.23 970 682 21 July 2008 MIT 65.69 37.83 440 970 3 May 2009 TAS_L 65.64 38.90 250 1160 28 August 2007 Land 2022, 11, 593 3 of 14 Table 1. Cont. Station Latitude Longitude Elevation Elevation-ELA Start Date Name ( N) ( W) (m) (m) TAS_U 65.70 38.87 570 840 15 August 2007 TAS_A 65.78 38.90 890 520 23 August 2013 QAS_L 61.03 46.85 280 1190 24 August 2007 QAS_M 61.10 46.83 630 840 11 August 2016 QAS_U 61.18 46.82 900 570 7 August 2008 QAS_A 61.24 46.73 1000 470 20 August 2012 NUK_L 64.48 49.54 530 885 20 August 2007 NUK_U 64.51 49.27 1120 295 20 August 2007 NUK_N 64.95 49.89 920 495 25 July 2010 KAN_L 67.10 49.95 670 745 1 September 2008 KAN_M 67.07 48.84 1270 145 2 September 2008 KAN_U 67.00 47.03 1840 425 4 April 2009 UPE_L 72.89 54.30 220 843 17 August 2009 UPE_U 72.89 53.58 940 123 17 August 2009 THU_L 76.40 68.27 570 493 9 August 2010 THU_U 76.42 68.15 760 303 9 August 2010 Land 2021, 10, x FOR PEER REVIEW 3 of 16 CEN 77.17 61.11 1880 817 23 May 2017 * L: Lower station, M: Middle station, U: Upper station. Figure 1. Map of Greenland showing the elevation and locations of the PROMICE AWS. Six tran- Figure 1. Map of Greenland showing the elevation and locations of the PROMICE AWS. Six transects sects are magnified in subplots. are magnified in subplots. 2.2. MODIS IST Products Table 1. Geographical characteristics of the meteorological stations used in this study. MODIS is a polar-orbiting, 36-channel, and across-track scanning spectro-radiometer Station Elevation- whose images cover Latituthe de (whole ° N) Longi region tudof e (the ° W)planet Elevatevery ion (m) one to two days [44 S]. taMODIS rt Date has Name ELA (m) been flying aboard NASA Earth Observation System (EOS) Terra satellites since 2000 and KPC_L * 79.91 24.08 370 −707 17 July 2008 Aqua since 2002 [45], so the GrIS swath-based daily gridded products began to be available KPC_U 79.83 25.17 870 −207 17 July 2008 at the beginning of 2000. Terra’s orbit around the Earth is timed so that it passes across EGP 75.62 35.97 2660 1583 1 May 2016 Greenland in the afternoon (14:30–16:30 UTC) [41]. MODIS Collection 6.1 IST used in this SCO_L 72.22 26.82 460 −1192 21 July 2008 study is from multilayer, daily IST-albedo-water vapor products developed using standard MODIS SCO_ datasets U from 72.3the 9 Terra satellite 27.23 [28]. 970 −682 21 July 2008 MIT 65.69 37.83 440 −970 3 May 2009 TAS_L 65.64 38.90 250 −1160 28 August 2007 TAS_U 65.70 38.87 570 −840 15 August 2007 TAS_A 65.78 38.90 890 −520 23 August 2013 QAS_L 61.03 46.85 280 −1190 24 August 2007 QAS_M 61.10 46.83 630 −840 11 August 2016 QAS_U 61.18 46.82 900 −570 7 August 2008 QAS_A 61.24 46.73 1000 −470 20 August 2012 NUK_L 64.48 49.54 530 −885 20 August 2007 NUK_U 64.51 49.27 1120 −295 20 August 2007 NUK_N 64.95 49.89 920 −495 25 July 2010 KAN_L 67.10 49.95 670 −745 1 September 2008 KAN_M 67.07 48.84 1270 −145 2 September 2008 KAN_U 67.00 47.03 1840 425 4 April 2009 UPE_L 72.89 54.30 220 −843 17 August 2009 UPE_U 72.89 53.58 940 −123 17 August 2009 THU_L 76.40 68.27 570 −493 9 August 2010 THU_U 76.42 68.15 760 −303 9 August 2010 CEN 77.17 61.11 1880 817 23 May 2017 * L: Lower station, M: Middle station, U: Upper station. Land 2022, 11, 593 4 of 14 Its algorithm was derived from the algorithm developed for the MODIS sea ice product MOD29 or MYD29 (MOD and MYD stand for Terra and Aqua products, respec- tively) [45–47]. The “MOD29” IST products during March 2000~December 2019, provide daily and monthly mean ISTs with polar-stereographic grids at 0.78  0.78 km resolu- tion [28]. As the input product of the MOD29 algorithm, MOD35 determines the cloud obscuration of the MODIS IST products. The cloudy grid cells without IST value are excluded calculation process. 2.3. MODIS Products Preprocessing The cloud coverage affects the accuracy of MODIS IST products. The internal cloud cover incorrectly identified instances of the ice surface as cloudless on a lot of days in JJA (June, July, and August) [8]. Therefore, the ISTs are actually the cloud top temperatures, far below the clear surface temperature (up to 30 C) [8]. Therefore, we checked the MODSI IST of each grid over two standard deviations to reduce outliers caused by cloud masks or other reasons. The percentage of the removed pixels was shown in Table 2. Table 2. The percentage of the MODIS IST pixels that had been removed. Year Percentage (%) Year Percentage (%) 2000 0.56 2010 0.46 2001 0.55 2011 0.49 2002 0.52 2012 0.48 2003 0.52 2013 0.50 2004 0.52 2014 0.50 2005 0.49 2015 0.50 2006 0.48 2016 0.47 2007 0.50 2017 0.48 2008 0.50 2018 0.48 2009 0.48 2019 0.45 As the ice/snow surface cannot exist in the form of ice/snow above 0 C [8], the fake pixels values (0 C) found on the ice sheet are removed. The most likely reason for a few pixels with the wrong IST is incorrect cloud masking caused by MODIS cloud mask [48]. Additionally, some pixel values representing no data, cloud cover, and fill value, are also removed. 2.4. Statistical Indexes for MODIS IST Assessment The daily MODIS ISTs used to make a comparison with the observations were extracted respectively from the nearest gridded MODIS IST at the ice sheet. Meanwhile, three classical statistical indexes were used to estimate the performance of the MODIS product, including the bias, root mean square (RMSE), and the correlation coefficient (R). The calculation formulas are as follows: (M O ) i i i=1 Bias = (1) (O M ) i i i=1 RMSE = (2) å (O O)(M M) i i i=1 s s r = (3) N N 2 2 (O O) (M M) å å i i i=1 i=1 Land 2022, 11, 593 5 of 14 where N is the available number of the sample, O is the observed IST, and M is the corresponding MODIS IST. 2.5. Anomaly and Trend Analysis As the MODIS IST dataset of the GrIS used in this study is available since March 2000, the annual anomaly of 2000 was not calculated. For each grid, the annual IST anomalies were calculated using the monthly data, relative to the multi-year average annual IST from 2001 to 2019. The summer IST includes the year 2000. When seasonal and annual IST trends of each grid were calculated, only the grids with 20-year data are used to calculate. 3. Results 3.1. MODIS IST Performance To initially obtain the performance of MODIS IST, 12 AWSs are selected according to their location and relative data volume. Figure 2 shows that the MODIS IST agrees well with the observations, with the R > 0.84 (TAS_ L and QAS_L). A significant characteristic Land 2021, 10, x FOR PEER REVIEW 6 of 16 is MODIS IST has a cold deviation of 2–3 C. MODIS IST shows the best agreement with KPC_U AWS observation, with a low RMSE (RMSE = 3.71 C) and a high correlation (R = 0.97), whereas, generally poorer agreement with TAS_L and QAS_L. Figure 2. MODIS and observed daily IST, the red diagonal represents the 1:1 line. Statistics show Figure 2. MODIS and observed daily IST, the red diagonal represents the 1:1 line. Statistics show Correlation coefficient (R), root mean squared error (RMSE), and bias. Correlation coefficient (R), root mean squared error (RMSE), and bias. Figure 3 also suggests that the bias indicates the IST tends to underestimate surface Figure 3 also suggests that the bias indicates the IST tends to underestimate surface temperature (2–4 C). The R and RMSE indicate that MODIS IST is highly consistent with temperature (2–4 °C). The R and RMSE indicate that MODIS IST is highly consistent ground measurements. Spatially, the performance of MODIS IST in the north is better with ground measurements. Spatially, the performance of MODIS IST in the north is bet- than that in the south, with a higher average R (0.96) and lower average RMSE (4.19 C). ter than that in the south, with a higher average R (0.96) and lower average RMSE (4.19 Compared with other regions, the R and RMSE show a better accuracy of MODIS IST on °C). Compared with other regions, the R and RMSE show a better accuracy of MODIS the northwest, northeast, and central GrIS. In these areas, the R is close to 1 and the range IST on the northwest, northeast, and central GrIS. In these areas, the R is close to 1 and of RMSE is concentrated at 3.7–4.9 C, indicating a high correlation. The MODIS IST shows the range of RMSE is concentrated at 3.7–4.9 °C, indicating a high correlation. The the best agreement with observations in the northeast, with the highest average R (0.96) MODIS IST shows the best agreement with observations in the northeast, with the high- est average R (0.96) and the lowest average RMSE (3.93 °C), however, generally poorer agreement with AWSs observations in the southeast. Land 2022, 11, 593 6 of 14 Land 2021, 10, x FOR PEER REVIEW 7 of 16 and the lowest average RMSE (3.93 C), however, generally poorer agreement with AWSs observations in the southeast. Fig Figure ure 33. . Cor Corr relelation ation coeffi coefc fi ien cient t (R), (R), bibias, as, an and d roo root t m mean ean squa squar red ed er err ror or (RMS (RMSE) E) bet between ween da daily ily MODIS MODIS IST and observations from all AWSs (AWSs shown in bold are above the ELA). IST and observations from all AWSs (AWSs shown in bold are above the ELA). Figure 4 compares MODIS IST time series with AWSs to analyze the typical tem- Figure 4 compares MODIS IST time series with AWSs to analyze the typical temporal poral variation and (in) consistency of time series derived from MODIS IST. KPC_U, variation and (in) consistency of time series derived from MODIS IST. KPC_U, KAN_U, KAN_U, and EGP were selected for their high-continuity data. The MODIS IST perfor- and EGP were selected for their high-continuity data. The MODIS IST performance is good, mance is good, and the time series are consistent well with observations, only with a and the time series are consistent well with observations, only with a slight underestimation slight underestimation in summer. Notably, the cold bias of MODIS IST mainly occurs in Land 2021, 10, x FOR PEER REVIEW 8 of 16 in summer. Notably, the cold bias of MODIS IST mainly occurs in winter. The average cold winter. The average cold deviations in summer and winter are 2.04 °C and 4.35 °C, re- deviations in summer and winter are 2.04 C and 4.35 C, respectively. spectively. Figure 4. Time series of MODIS IST (red line) and surface temperature (green line) from EGP (a), Figure 4. Time series of MODIS IST (red line) and surface temperature (green line) from EGP (a), KAN_U KAN_U ((b b )), , aand nd KKPC_U PC_U (c). (c ). Overall, the performance of MODIS products is good, and the range of the cold bias is concentrated at 2–4 °C. A cold deviation of 0.98 °C was found when MODIS IST was compared with the observation data measured by Summit Station of Greenland [28]. The results of thermochron data on the Summit also showed a cold deviation of 3.14 °C of MODIS IST [45]. Given the limited coverage of AWSs on GrIS, we suggest that ISTs ob- tained from the MODIS IST products are characterized by high accuracy and could be reliably used as an alternative or supplement to Greenland temperature monitoring, es- pecially for IST in summer. Generally, fine space-based climate information obtained by analyzing MODIS IST can effectively evaluate the temporal and spatial variability of the GrIS and be applied in many fields. 3.2. Temporal and Spatial Variability of GrIS IST Figure 5 are clear-sky IST maps illustrating the MODIS IST on the GrIS at seasonal and annual scales. The MODIS IST revealed distinct spatial gradients varying from outer (low elevation) to inner (high elevation), showing a colder inside with higher elevation. Apparently, the IST in the northern GrIS is much lower than that in the southern region while there is little difference between the eastern and western GrIS. The mean autumn IST is a little colder than the spring and annual mainly reflected in the inner of the GrIS. The mean summer IST is mainly concentrated between −20 °C and −10 °C with the high- est IST (−1.33 °C) in the southwest margin of the GrIS and the lowest IST (−17.98 °C) in the inner. Conversely, the mean winter IST is mainly concentrated between −50 °C and −30 °C with the highest and lowest IST values of −9.85 °C and −50.13 °C, respectively. The minimum IST values in winter are found at high elevations just as in summer but Land 2022, 11, 593 7 of 14 Overall, the performance of MODIS products is good, and the range of the cold bias is concentrated at 2–4 C. A cold deviation of 0.98 C was found when MODIS IST was compared with the observation data measured by Summit Station of Greenland [28]. The results of thermochron data on the Summit also showed a cold deviation of 3.14 C of MODIS IST [45]. Given the limited coverage of AWSs on GrIS, we suggest that ISTs obtained from the MODIS IST products are characterized by high accuracy and could be reliably used as an alternative or supplement to Greenland temperature monitoring, especially for IST in summer. Generally, fine space-based climate information obtained by analyzing MODIS IST can effectively evaluate the temporal and spatial variability of the GrIS and be applied in many fields. 3.2. Temporal and Spatial Variability of GrIS IST Figure 5 are clear-sky IST maps illustrating the MODIS IST on the GrIS at seasonal and annual scales. The MODIS IST revealed distinct spatial gradients varying from outer (low elevation) to inner (high elevation), showing a colder inside with higher elevation. Apparently, the IST in the northern GrIS is much lower than that in the southern region while there is little difference between the eastern and western GrIS. The mean autumn IST is a little colder than the spring and annual mainly reflected in the inner of the GrIS. The mean summer IST is mainly concentrated between 20 C and 10 C with the highest IST (1.33 C) in the southwest margin of the GrIS and the lowest IST (17.98 C) Land 2021, 10, x FOR PEER REVIEW 9 of 16 in the inner. Conversely, the mean winter IST is mainly concentrated between 50 C and 30 C with the highest and lowest IST values of 9.85 C and 50.13 C, respectively. The minimum IST values in winter are found at high elevations just as in summer but tend to migrate farther north, indicating that the coldest ISTs are not always located at tend to migrate farther north, indicating that the coldest ISTs are not always located at the the highest elevations, consistent with the previous study [49]. highest elevations, consistent with the previous study [49]. Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter (d) (d) and the mean annual MODIS IST (e) from March 2000 to December 2019. and the mean annual MODIS IST (e) from March 2000 to December 2019. 3.3. GrIS IST Anomalies 3.3. GrIS IST Anomalies Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 re- respectively. The IST anomalies between 1 C and 1 C are dominated in most of the spectively. The IST anomalies between −1 °C and 1 °C are dominated in most of the study period except for some abnormally warm and cold years. The largest positive IST study period except for some abnormally warm and cold years. The largest positive IST anomalies (exceeds 3 C) occurred in the southwest in 2010 when almost the entire GrIS anomalies (exceeds 3 °C) occurred in the southwest in 2010 when almost the entire GrIS experienced positive IST anomalies consistent with [8]. In that year, Greenland experienced experienced positive IST anomalies consistent with [8]. In that year, Greenland experi- melting up to another 60 days compared with the 1960–2010 average (1980–2010 were enced melting up to another 60 days compared with the 1960–2010 average (1980–2010 simulated data), with the largest differences occurring at the southwestern and western were simulated data), with the largest differences occurring at the southwestern and margins of the GrIS [6]. The previous study shows that the unusual melt event in 2010 was western margins of the GrIS [6]. The previous study shows that the unusual melt event trigged by large positive air temperature anomalies during May, accelerating snowpack in 2010 was trigged by large positive air temperature anomalies during May, accelerat- metamorphism, and premature bare ice exposure [50]. The GrIS IST anomaly in May 2010 ing snowpack metamorphism, and premature bare ice exposure [50]. The GrIS IST may be relevant to the eruption of Eyjafjallajokull ˝ volcano on March 2010 which once had a anomaly in May 2010 may be relevant to the eruption of Eyjafjallajőkull volcano on profound influence on the environment and climatic conditions of neighboring regions [51]. March 2010 which once had a profound influence on the environment and climatic con- With the positive albedo feedback and the fact that the wet snow could absorb up to three ditions of neighboring regions [51]. With the positive albedo feedback and the fact that times more incident solar energy than dry snow, the further melt was fostered [52]. the wet snow could absorb up to three times more incident solar energy than dry snow, the further melt was fostered [52]. Figure 6. Annual IST anomalies (°C) calculated from monthly-derived MODIS IST. In addition to 2010, relatively large positive IST anomalies (1–3 °C) were mainly found in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found in central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire GrIS in 2016 while northern and western GrIS in 2019. By comparison, a relatively large Land 2021, 10, x FOR PEER REVIEW 9 of 16 tend to migrate farther north, indicating that the coldest ISTs are not always located at the highest elevations, consistent with the previous study [49]. Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter (d) and the mean annual MODIS IST (e) from March 2000 to December 2019. 3.3. GrIS IST Anomalies Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 re- spectively. The IST anomalies between −1 °C and 1 °C are dominated in most of the study period except for some abnormally warm and cold years. The largest positive IST anomalies (exceeds 3 °C) occurred in the southwest in 2010 when almost the entire GrIS experienced positive IST anomalies consistent with [8]. In that year, Greenland experi- enced melting up to another 60 days compared with the 1960–2010 average (1980–2010 were simulated data), with the largest differences occurring at the southwestern and western margins of the GrIS [6]. The previous study shows that the unusual melt event in 2010 was trigged by large positive air temperature anomalies during May, accelerat- ing snowpack metamorphism, and premature bare ice exposure [50]. The GrIS IST anomaly in May 2010 may be relevant to the eruption of Eyjafjallajőkull volcano on March 2010 which once had a profound influence on the environment and climatic con- ditions of neighboring regions [51]. With the positive albedo feedback and the fact that Land 2022, 11, 593 the wet snow could absorb up to three times more incident solar energy than d 8 of ry 14 snow, the further melt was fostered [52]. Land 2021, 10, x FOR PEER REVIEW 10 of 16 Figure 6. Annual IST anomalies ( C) calculated from monthly-derived MODIS IST. Figure 6. Annual IST anomalies (°C) calculated from monthly-derived MODIS IST. In addition to 2010, relatively large positive IST anomalies (1–3 C) were mainly found In addition to 2010, relatively large positive IST anomalies (1–3 °C) were mainly region of negative IST anomalies that between –1 °C and –3 °C were found in 2001, 2011, in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found in found in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire GrIS in 2013 and 2015. Especially, the largest negative IST anomalies (below −3 °C) was found in in central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire 2016 while northern and western GrIS in 2019. By comparison, a relatively large region of 2015 in southwestern GrIS. GrIS negative in 201IST 6 wh anomalies ile north that ern between and western –1 C G and rIS–3 in C 20wer 19. eBfound y comin pa 2001, rison 2011, , a re2013 latively and large Figure 7 shows the temporal and spatial characteristics of the mean summer IST 2015. Especially, the largest negative IST anomalies (below 3 C) was found in 2015 in anomaly from 2000 to 2019, respectively. As reported by Hall et al. [8], the summer of southwestern GrIS. 2012 is the warmest summer, with the southeast experiencing the largest positive IST Figure 7 shows the temporal and spatial characteristics of the mean summer IST anom anomaly alies (ex from ceed 2000 ing to 3 2019, °C)r,espectively related to . As th re eported strongest by Hall No etrth al. [A 8],m the erica summer n hea of t 2012 wave since is the warmest summer, with the southeast experiencing the largest positive IST anomalies 1895 [53]. In 2012, almost the entire GrIS (98.6%) experienced an extreme melting event., (exceeding 3 C), related to the strongest North American heat wave since 1895 [53]. In 2012, even at Summit [7]. This extreme melt event was associated with an abnormal ridge of almost the entire GrIS (98.6%) experienced an extreme melting event., even at Summit [7]. warm air that was stagnant over Greenland and the radiative effects of low-level liquid This extreme melt event was associated with an abnormal ridge of warm air that was layer clouds [7,54]. Previous studies had shown that on 11–13 July, the GrIS experienced stagnant over Greenland and the radiative effects of low-level liquid layer clouds [7,54]. the most extensive melt extent based on satellite-derived data [7,9]. Figure 8 are the Previous studies had shown that on 11–13 July, the GrIS experienced the most extensive MOD melt IS extent IST on based 11–1 on 3 Jsatellite-derived uly of the GrIS data sho [7 wi ,9]. ng Figur thae t 8 th ar ee IS the T MODIS is conce IST ntra onted 11–13 beJuly tween 0 °C of the GrIS showing that the IST is concentrated between 0 C and 3 C, especially the and −3 °C, especially the western GrIS experienced the largest area of highest IST. western GrIS experienced the largest area of highest IST. Figure 7. Summer IST anomalies ( C) calculated from monthly IST (2000 to 2019). Figure 7. Summer IST anomalies (°C) calculated from monthly IST (2000 to 2019). Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is under cloud cover or has no data is considered as no data. In addition, a relatively large area of positive IST anomalies (0–2 °C) in summer was mainly found in 2007, 2010, 2016 and 2019. The recent warm summer found in our study period is the summer of 2019, agreeing with the previous study [55,56]. The increasing Land 2021, 10, x FOR PEER REVIEW 10 of 16 region of negative IST anomalies that between –1 °C and –3 °C were found in 2001, 2011, 2013 and 2015. Especially, the largest negative IST anomalies (below −3 °C) was found in 2015 in southwestern GrIS. Figure 7 shows the temporal and spatial characteristics of the mean summer IST anomaly from 2000 to 2019, respectively. As reported by Hall et al. [8], the summer of 2012 is the warmest summer, with the southeast experiencing the largest positive IST anomalies (exceeding 3 °C), related to the strongest North American heat wave since 1895 [53]. In 2012, almost the entire GrIS (98.6%) experienced an extreme melting event., even at Summit [7]. This extreme melt event was associated with an abnormal ridge of warm air that was stagnant over Greenland and the radiative effects of low-level liquid layer clouds [7,54]. Previous studies had shown that on 11–13 July, the GrIS experienced the most extensive melt extent based on satellite-derived data [7,9]. Figure 8 are the MODIS IST on 11–13 July of the GrIS showing that the IST is concentrated between 0 °C and −3 °C, especially the western GrIS experienced the largest area of highest IST. Land 2022, 11, 593 9 of 14 In addition, a relatively large area of positive IST anomalies (0–2 C) in summer was mainly found in 2007, 2010, 2016 and 2019. The recent warm summer found in our study period is the summer of 2019, agreeing with the previous study [55,56]. The increasing IST is the basic factor causing the increasing GrIS melting [57]. The warm, moist air intrusions could potentially drive melt events [54,58]. At the end of July 2019, the GrIS melting extent achieved more than 62%, even at Summit [55] and the ablation zones continued through August and September, entering accumulation conditions in early October [59]. Contrary to the above-average warmth conditions during the summers mentioned above, relatively Figure 7. Summer IST anomalies (°C) calculated from monthly IST (2000 to 2019). large areas of negative IST anomalies were found in 2000, 2001, 2002, 2006, 2009, 2013, 2017 and 2018. Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is under cloud cover or has no data is considered as no data. under cloud cover or has no data is considered as no data. 3.4. IST Trend Analysis of GrIS Figure 9 shows the positive IST trends mainly in spring and summer and negative In addition, a relatively large area of positive IST anomalies (0–2 °C) in summer was trends in fall and winter. The annual IST trends are much more moderate than the seasonal mainly found in 200IST 7, 2 tr0 ends 10, and 201the 6 amost nd 2IST 019 tr.ends The arre e concentrated cent warm ~+ sum 0.5mC/decade, er found showing in our ast lowe udy r spatial variability. On an annual scale, the northern GrIS mainly experienced positive IST period is the summer of 2019, agreeing with the previous study [55,56]. The increasing trends (0~0.5 C/decade) while the southern GrIS mainly experienced negative IST trends (1~0 C/decade). Autumn cooling was common and Hall et al. [8] also found a decrease in IST (~1.49  1.20 C/decade) in autumn during the period 2000–2012. Similarly, this work also showed the northwestern GrIS experienced positive IST trends except autumn during the study period, especially in spring with the positive IST trends reaching up to 1.5 C/decade. Furthermore, the minimum negative IST trends (below 1.5 C/decade) were found in the small area of southern and eastern GrIS in winter and a small area of southern GrIS in autumn. Land 2021, 10, x FOR PEER REVIEW 11 of 16 IST is the basic factor causing the increasing GrIS melting.[57]. The warm, moist air in- trusions could potentially drive melt events [54,58]. At the end of July 2019, the GrIS melting extent achieved more than 62%, even at Summit [55] and the ablation zones con- tinued through August and September, entering accumulation conditions in early Octo- ber [59]. Contrary to the above-average warmth conditions during the summers men- tioned above, relatively large areas of negative IST anomalies were found in 2000, 2001, 2002, 2006, 2009, 2013, 2017 and 2018. 3.4. IST Trend Analysis of GrIS Figure 9 shows the positive IST trends mainly in spring and summer and negative trends in fall and winter. The annual IST trends are much more moderate than the sea- sonal IST trends and the most IST trends are concentrated ~+−0.5 °C/decade, showing a lower spatial variability. On an annual scale, the northern GrIS mainly experienced posi- tive IST trends (0~0.5 °C/decade) while the southern GrIS mainly experienced negative IST trends (−1~0 °C/decade). Autumn cooling was common and Hall et al. [8] also found a decrease in IST (~−1.49 ± 1.20 °C/decade) in autumn during the period 2000–2012. Simi- larly, this work also showed the northwestern GrIS experienced positive IST trends ex- cept autumn during the study period, especially in spring with the positive IST trends reaching up to 1.5 °C/decade. Furthermore, the minimum negative IST trends (below −1.5 °C/decade) were found in the small area of southern and eastern GrIS in winter and Land 2022, 11, 593 10 of 14 a small area of southern GrIS in autumn. Figure 9. Seasonal MODIS IST trends over the GrIS in spring (a), summer (b), autumn (c) and winter Figure 9. Seasonal MODIS IST trends over the GrIS in spring (a), summer (b), autumn (c) and win- (d) from March 2000 to December 2019 and the annual MODIS IST (e). ter (d) from March 2000 to December 2019 and the annual MODIS IST (e). 4. Discussion 4. Discussion GrIS melt and runoff have increased rapidly since the early 1990s [26]. The low density GrIS melt and runoff have increased rapidly since the early 1990s [26]. The low and discontinuity of AWS measurement are the inherent limitation in evaluating the melt dens onity the aentir nd deisice con sheet. tinuity Surface of AWS melting measurement conditionsare that thcannot e inherbe ent measur limitaed tiodue n in to evthe aluating AWSs constraints can only be obtained by remote sensing [60]. Given the significance of the melt on the entire ice sheet. Surface melting conditions that cannot be measured due remotely sensed melt products for monitoring GrIS melt, an assessment of the MODIS to the AWSs constraints can only be obtained by remote sensing [60]. Given the signifi- IST-derived melt data is essential. cance of remotely sensed melt products for monitoring GrIS melt, an assessment of the The MODIS IST-derived melt data is determined by the melt threshold (1 C) and MODIS IST-derived melt data is essential. a non-cloud-obscured IST grid that is 1 C is defined as “melt”. Therefore, the melt The MODIS IST-derived melt data is determined by the melt threshold (−1 °C) and threshold is the key factor to be validated. The IST products have an accuracy of 1 C a non-cloud-obscured IST grid that is ≥−1 °C is defined as “melt”. Therefore, the melt and melt may occur when temperatures are slightly below freezing with a strong solar threshold is the key factor to be validated. The IST products have an accuracy of ±1 °C radiation [28]. Given the completeness and amount of the AWSs daily ablation data, the daily ablation data from 5 AWSs at low elevations and 3 AWSs at high elevations were and melt may occur when temperatures are slightly below freezing with a strong solar selected to validate the melt threshold. The daily ablation data was set to 25 mm based radiation [28]. Given the completeness and amount of the AWSs daily ablation data, the on the accuracy of the pressure transducer [37]. Only the data that were both available daily ablation data from 5 AWSs at low elevations and 3 AWSs at high elevations were for MODIS and AWS were used in this study. The days that ware detected melt or no melt both from MODIS and AWS were recorded as “BB” and these days can be considered well-matched days. If the days were detected melt only from MODIS or AWS, then they were recorded as “MM” or “SM” and these days can be considered failed-matched days. Figure 10 shows the day-to-day melt agreement conditions between MODIS and AWSs. Obviously, AWSs at low elevations reproduced more melt occasions than at high elevations. In the strong ablation year of 2012, except for the KPC_L, MODIS IST reproduced fewer melt conditions than AWSs. In general, MODIS IST performed an average agreement of 55% with the highest agreement in KPC_L (61%), located at the northeastern GrIS. NUK_L at southwestern GrIS had the lowest consistency of 41% with much fewer melt days than that detected by MODIS IST. Other AWSs show an agreement between 50% and 60%. Overall, the 1 C melt threshold suggested a 41–61% consistency of melt conditions (BB) between MODIS and AWSs, and generally, fewer MODIS melt days occurred than AWSs. Given the fewer MODIS melt days based on 1 C and the cold bias found in this work, lower melt thresholds were used to possibly achieve higher consistencies (Figure 11). The results showed that consistencies were 46–66% and 51–70% based on melt thresholds respectively 1.5 C and 2 C, respectively. With the melt threshold decreasing, the consistency was generally becoming higher especially for the stations QAS_L, NUK_L, KAN_L, and TAS_U while KPC_L and QAS_U had lower consistencies. Land 2021, 10, x FOR PEER REVIEW 12 of 16 selected to validate the melt threshold. The daily ablation data was set to 25 mm based on the accuracy of the pressure transducer [37]. Only the data that were both available for MODIS and AWS were used in this study. The days that ware detected melt or no melt both from MODIS and AWS were recorded as “BB” and these days can be consid- ered well-matched days. If the days were detected melt only from MODIS or AWS, then they were recorded as “MM” or “SM” and these days can be considered failed-matched days. Figure 10 shows the day-to-day melt agreement conditions between MODIS and AWSs. Obviously, AWSs at low elevations reproduced more melt occasions than at high elevations. In the strong ablation year of 2012, except for the KPC_L, MODIS IST repro- duced fewer melt conditions than AWSs. In general, MODIS IST performed an average agreement of 55% with the highest agreement in KPC_L (61%), located at the northeast- ern GrIS. NUK_L at southwestern GrIS had the lowest consistency of 41% with much fewer melt days than that detected by MODIS IST. Other AWSs show an agreement be- tween 50% and 60%. Overall, the −1 °C melt threshold suggested a 41–61% consistency of melt conditions (BB) between MODIS and AWSs, and generally, fewer MODIS melt days occurred than AWSs. Given the fewer MODIS melt days based on −1 °C and the cold bias found in this work, lower melt thresholds were used to possibly achieve higher consistencies (Figure 11). The results showed that consistencies were 46–66% and 51– 70% based on melt thresholds respectively −1.5 °C and −2 °C, respectively. With the melt threshold decreasing, the consistency was generally becoming higher especially for the Land 2022, 11, 593 11 of 14 stations QAS_L, NUK_L, KAN_L, and TAS_U while KPC_L and QAS_U had lower con- sistencies. Land 2021, 10, x FOR PEER REVIEW 13 of 16 Fig Figure ure 1 10. 0. Th The e c consistency onsistency b between etween MODI MODIS S a and nd A AWSs WSs melt melt condition. condition. Figure Figure 11. 11. Consistency Consistency of of melt melt conditions conditions (BB) (BB) between between MODIS MODIS and and station station based based on on the the melt melt threshold −1 °C, −1.5 °C, and −2 °C. threshold 1 C, 1.5 C, and 2 C. 5. Conclusions 5. Conclusions In this study, based on PROMICE AWSs observations, three statistical indexes, bias, In this study, based on PROMICE AWSs observations, three statistical indexes, bias, RMSE, and R are used to comprehensively evaluate the performance of the MODIS IST over RMSE, and R are used to comprehensively evaluate the performance of the MODIS IST the GrIS. Generally, MODIS IST has a widespread cold bias over the GrIS, ranging from 2 over the GrIS. Generally, MODIS IST has a widespread cold bias over the GrIS, ranging from −2 to −4 °C and this cold bias mainly occurred in winter. The consistency between MODIS IST and in-situ IST is very high on the GrIS, with an average of R = 0.92 and RMSE = 4.36 °C. Spatially, the R and RMSE agree on a better accuracy of MODIS IST on the northwestern, northeastern, and central GrIS. The MODIS IST can well demonstrate the temporal and spatial variability of GrIS. The mean summer IST is mainly concentrated between −20 °C and −10 °C with the high- est and lowest IST values of −1.33 °C and −17.98 °C, respectively. Besides, the largest positive IST anomalies (exceeding 3 °C) occurred in southwestern GrIS in 2010. Summer 2012 was the warmest summer with the largest positive IST anomalies (exceeding 3 °C) in the southeastern GrIS. The positive IST trends are mainly in spring and summer and the negative IST trends are in autumn and winter. Importantly, most of the annual IST trends are concentrated between ~±0.5 °C/decade. This work also verifies the rationality of the MODIS IST melt threshold of −1 °C us- ing the daily ablation data from AWSs. The results show that the −1 °C melt threshold suggested a 41–61% consistency of melt conditions between MODIS and AWSs. Fur- thermore, with the melt threshold decreasing (−1.5 °C, −2 °C), the consistency is general- ly becoming higher. Author Contributions: X.Y.: Formal analysis, visualization, writing—original draft. T.W.: formal analysis, visualization, writing—original draft. M.D.: formal analysis, writing—original draft. Y.W.: formal analysis, methodology. W.S.: formal analysis, methodology. Q.Z.: methodology, software, validation. B.H.: funding acquisition, supervision, conceptualization, methodology, val- idation, writing—review and editing. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Natural Science Foundation of China grant number 42171121, 41701059. And The APC was funded by 42171121. Data Availability Statement: All data and the script of the whole processes are available through an email request to the authors. Land 2022, 11, 593 12 of 14 to 4 C and this cold bias mainly occurred in winter. The consistency between MODIS IST and in-situ IST is very high on the GrIS, with an average of R = 0.92 and RMSE = 4.36 C. Spatially, the R and RMSE agree on a better accuracy of MODIS IST on the northwestern, northeastern, and central GrIS. The MODIS IST can well demonstrate the temporal and spatial variability of GrIS. The mean summer IST is mainly concentrated between 20 C and 10 C with the highest and lowest IST values of 1.33 C and 17.98 C, respectively. Besides, the largest positive IST anomalies (exceeding 3 C) occurred in southwestern GrIS in 2010. Summer 2012 was the warmest summer with the largest positive IST anomalies (exceeding 3 C) in the southeastern GrIS. The positive IST trends are mainly in spring and summer and the negative IST trends are in autumn and winter. Importantly, most of the annual IST trends are concentrated between ~0.5 C/decade. This work also verifies the rationality of the MODIS IST melt threshold of 1 C using the daily ablation data from AWSs. The results show that the 1 C melt threshold suggested a 41–61% consistency of melt conditions between MODIS and AWSs. Further- more, with the melt threshold decreasing (1.5 C, 2 C), the consistency is generally becoming higher. Author Contributions: X.Y.: Formal analysis, visualization, writing—original draft. T.W.: formal analysis, visualization, writing—original draft. M.D.: formal analysis, writing—original draft. Y.W.: formal analysis, methodology. W.S.: formal analysis, methodology. Q.Z.: methodology, software, vali- dation. B.H.: funding acquisition, supervision, conceptualization, methodology, validation, writing— review and editing. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Natural Science Foundation of China grant number 42171121, 41701059. And The APC was funded by 42171121. Data Availability Statement: All data and the script of the whole processes are available through an email request to the authors. Acknowledgments: This work was funded by the Natural Science Foundation of China (42171121, 41701059). The authors gratefully acknowledge data availability from the Programme for Monitoring of the Greenland Ice Sheet (PROMICE) and the Greenland Analogue Project (GAP) provided by the Geological Survey of Denmark and Greenland (GEUS) at http://www.promice.dk (accessed on 6 October 2021). Conflicts of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References 1. Steffen, K.; Box, J. Surface climatology of the Greenland ice sheet: Greenland climate network 1995–1999. J. Geophys. Res.-Atmos. 2001, 106, 33951–33964. [CrossRef] 2. Mouginot, J.; Rignot, E.; Bjork, A.A.; van den Broeke, M.; Millan, R.; Morlighem, M.; Noel, B.; Scheuchl, B.; Wood, M. Forty-six years of Greenland Ice Sheet mass balance from 1972 to 2018. Proc. Natl. Acad. Sci. USA 2019, 116, 9239–9244. [CrossRef] [PubMed] 3. Blunden, J.; Arndt, D. State of the Climate in 2019. Bull. Am. Meteorol. Soc. 2020, 101, S1–S429. [CrossRef] 4. Hanna, E.; Huybrechts, P.; Steffen, K.; Cappelen, J.; Huff, R.; Shuman, C.; Irvine-Fynn, T.; Wise, S.; Griffiths, M. Increased runoff from melt from the Greenland Ice Sheet: A response to global warming. J. Clim. 2008, 21, 331–341. [CrossRef] 5. Tedesco, M.; Serreze, M.; Fettweis, X. Diagnosing the extreme surface melt event over southwestern Greenland in 2007. Cryosphere 2008, 2, 159–166. [CrossRef] 6. Mernild, S.H.; Mote, T.L.; Liston, G.E. Greenland ice sheet surface melt extent and trends: 1960–2010. J. Glaciol. 2011, 57, 621–628. [CrossRef] 7. Nghiem, S.V.; Hall, D.K.; Mote, T.L.; Tedesco, M.; Albert, M.R.; Keegan, K.; Shuman, C.A.; DiGirolamo, N.E.; Neumann, G. The extreme melt across the Greenland ice sheet in 2012. Geophys. Res. Lett. 2012, 39, L20502. [CrossRef] 8. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Box, J.E.; Koenig, L.S. Variability in the surface temperature and melt extent of the Greenland ice sheet from MODIS. Geophys. Res. Lett. 2013, 40, 2114–2120. [CrossRef] 9. Hanna, E.; Fettweis, X.; Mernild, S.H.; Cappelen, J.; Ribergaard, M.H.; Shuman, C.A.; Steffen, K.; Wood, L.; Mote, T.L. Atmospheric and oceanic climate forcing of the exceptional Greenland ice sheet surface melt in summer 2012. Int. J. Climatol. 2014, 34, 1022–1037. [CrossRef] Land 2022, 11, 593 13 of 14 10. Rignot, E.; Velicogna, I.; van den Broeke, M.R.; Monaghan, A.; Lenaerts, J. Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophys. Res. Lett. 2011, 38, L05503. [CrossRef] 11. Enderlin, E.M.; Howat, I.M. Submarine melt rate estimates for floating termini of Greenland outlet glaciers (2000–2010). J. Glaciol. 2013, 59, 67–75. [CrossRef] 12. Fettweis, X.; Hanna, E.; Lang, C.; Belleflamme, A.; Erpicum, M.; Gallee, H. Important role of the mid-tropospheric atmospheric circulation in the recent surface melt increase over the Greenland ice sheet. Cryosphere 2013, 7, 241–248. [CrossRef] 13. Wouters, B.; Bamber, J.L.; van den Broeke, M.R.; Lenaerts, J.T.M.; Sasgen, I. Limits in detecting acceleration of ice sheet mass loss due to climate variability. Nat. Geosci. 2013, 6, 613–616. [CrossRef] 14. Comiso, J.C.; Yang, J.; Honjo, S.; Krishfield, R.A. Detection of change in the Arctic using satellite and in situ data. J. Geophys. Res. Ocean. 2003, 108, C12. [CrossRef] 15. Wang, X.J.; Key, J.R. Recent trends in arctic surface, cloud, and radiation properties from space. Science 2003, 299, 1725–1728. [CrossRef] 16. Wang, X.J.; Key, J.R. Arctic surface, cloud, and radiation properties based on the AVHRR Polar Pathfinder dataset. Part I: Spatial and temporal characteristics. J. Clim. 2005, 18, 2558–2574. [CrossRef] 17. Wang, X.J.; Key, J.R. Arctic surface, cloud, and radiation properties based on the AVHRR Polar Pathfinder dataset. Part II: Recent trends. J. Clim. 2005, 18, 2575–2593. [CrossRef] 18. Mote, T.L. Greenland surface melt trends 1973–2007: Evidence of a large increase in 2007. Geophys. Res. Lett. 2007, 34, L22507. [CrossRef] 19. van den Broeke, M.; Bamber, J.; Ettema, J.; Rignot, E.; Schrama, E.; van de Berg, W.J.; van Meijgaard, E.; Velicogna, I.; Wouters, B. Partitioning Recent Greenland Mass Loss. Science 2009, 326, 984–986. [CrossRef] 20. Reddy, S. Comments on IPCC’s 24th September 2019 Report on “The Ocean and Cryosphere in a Changing Climate: Summary for Policy Makers”. Acta Sci. Agric 2019, 3, 16–19. [CrossRef] 21. Pörtner, H.-O.; Roberts, D.C.; Masson-Delmotte, V.; Zhai, P.; Tignor, M.; Poloczanska, E.; Weyer, N. The ocean and cryosphere in a changing climate. IPCC Spec. Rep. Ocean Cryosphere A Chang. Clim. 2019. Available online: https://www.ipcc.ch/site/assets/ uploads/sites/3/2019/12/02_SROCC_FM_FINAL.pdf (accessed on 15 March 2022). 22. Muntjewerf, L.; Petrini, M.; Vizcaino, M.; da Silva, C.E.; Sellevold, R.; Scherrenberg, M.D.W.; Thayer-Calder, K.; Bradley, S.L.; Lenaerts, J.T.M.; Lipscomb, W.H.; et al. Greenland Ice Sheet Contribution to 21st Century Sea Level Rise as Simulated by the Coupled CESM2.1-CISM2.1. Geophys. Res. Lett. 2020, 47, e2019GL086836. [CrossRef] 23. Calov, R.; Beyer, S.; Greve, R.; Beckmann, J.; Willeit, M.; Kleiner, T.; Rueckamp, M.; Humbert, A.; Ganopolski, A. Simulation of the future sea level contribution of Greenland with a new glacial system model. Cryosphere 2018, 12, 3097–3121. [CrossRef] 24. Rueckamp, M.; Greve, R.; Humbert, A. Comparative simulations of the evolution of the Greenland ice sheet under simplified Paris Agreement scenarios with the models SICOPOLIS and ISSM. Polar Sci. 2019, 21, 14–25. [CrossRef] 25. Bell, R.E. The role of subglacial water in ice-sheet mass balance. Nat. Geosci. 2008, 1, 297–304. [CrossRef] 26. van den Broeke, M.R.; Enderlin, E.M.; Howat, I.M.; Munneke, P.K.; Noel, B.P.Y.; van de Berg, W.J.; van Meijgaard, E.; Wouters, B. On the recent contribution of the Greenland ice sheet to sea level change. Cryosphere 2016, 10, 1933–1946. [CrossRef] 27. Smith, L.C.; Chu, V.W.; Yang, K.; Gleason, C.J.; Pitcher, L.H.; Rennermalm, A.K.; Legleiter, C.J.; Behar, A.E.; Overstreet, B.T.; Moustafa, S.E.; et al. Efficient meltwater drainage through supraglacial streams and rivers on the southwest Greenland ice sheet. Proc. Natl. Acad. Sci. USA 2015, 112, 1001–1006. [CrossRef] 28. Hall, D.K.; Cullather, R.I.; DiGirolamo, N.E.; Comiso, J.C.; Medley, B.C.; Nowicki, S.M. A Multilayer Surface Temperature, Surface Albedo, and Water Vapor Product of Greenland from MODIS. Remote Sens. 2018, 10, 555. [CrossRef] 29. Miller, N.B.; Shupe, M.D.; Cox, C.J.; Noone, D.; Persson, P.O.G.; Steffen, K. Surface energy budget responses to radiative forcing at Summit, Greenland. Cryosphere 2017, 11, 497–516. [CrossRef] 30. Noel, B.; van de Berg, W.J.; van Wessem, J.M.; van Meijgaard, E.; van As, D.; Lenaerts, J.T.M.; Lhermitte, S.; Munneke, P.K.; Smeets, C.J.P.P.; van Ulft, L.H.; et al. Modelling the climate and surface mass balance of polar ice sheets using RACMO2-Part 1: Greenland (1958–2016). Cryosphere 2018, 12, 811–831. [CrossRef] 31. Noel, B.; van de Berg, W.J.; Lhermitte, S.; van den Broeke, M.R. Rapid ablation zone expansion amplifies north Greenland mass loss. Sci. Adv. 2019, 5, eaaw0123. [CrossRef] [PubMed] 32. Huai, B.; van den Broeke, M.R.; Reijmer, C.H. Long-term surface energy balance of the western Greenland Ice Sheet and the role of large-scale circulation variability. Cryosphere 2020, 14, 4181–4199. [CrossRef] 33. Nielsen-Englyst, P.; Hoyer, J.L.; Madsen, K.S.; Tonboe, R.T.; Dybkjaer, G.; Skarpalezos, S. Deriving Arctic 2 m air temperatures over snow and ice from satellite surface temperature measurements. Cryosphere 2021, 15, 3035–3057. [CrossRef] 34. Shuman, C.A.; Steffen, K.; Box, J.E.; Stearns, C.R. A dozen years of temperature observations at the summit: Central Greenland automatic weather stations 1987–1999. J. Appl. Meteorol. 2001, 40, 741–752. [CrossRef] 35. Ahlstrom, A.P.; Gravesen, P.; Andersen, S.B.; van As, D.; Citterio, M.; Fausto, R.S.; Nielsen, S.; Jepsen, H.F.; Kristensen, S.S.; Christensen, E.L.; et al. A new programme for monitoring the mass loss of the Greenland ice sheet. Geol. Surv. Den. Greenl. Bull. 2008, 15, 61–64. [CrossRef] 36. van As, D.; Fausto, R.S.; Team, P.P. Programme for Monitoring of the Greenland Ice Sheet (PROMICE): First temperature and ablation records. Geol. Surv. Den. Greenl. Bull. 2011, 23, 73–76. [CrossRef] Land 2022, 11, 593 14 of 14 37. Fausto, R.S.; Van As, D.; Ahlstrom, A.P.; Citterio, M. Assessing the accuracy of Greenland ice sheet ice ablation measurements by pressure transducer. J. Glaciol. 2012, 58, 1144–1150. [CrossRef] 38. Box, J.E.; Rinke, A. Evaluation of Greenland ice sheet surface climate in the HIRHAM regional climate model using automatic weather station data. J. Clim. 2003, 16, 1302–1319. [CrossRef] 39. Leeson, A.A.; Van Wessem, J.M.; Ligtenberg, S.R.M.; Shepherd, A.; Van Den Broeke, M.R.; Killick, R.; Skvarca, P.; Marinsek, S.; Colwell, S. Regional climate of the Larsen B embayment 1980–2014. J. Glaciol. 2017, 63, 683–690. [CrossRef] 40. Medley, B.; Joughin, I.; Das, S.B.; Steig, E.J.; Conway, H.; Gogineni, S.; Criscitiello, A.S.; McConnell, J.R.; Smith, B.E.; van den Broeke, M.R.; et al. Airborne-radar and ice-core observations of annual snow accumulation over Thwaites Glacier, West Antarctica confirm the spatiotemporal variability of global and regional atmospheric models. Geophys. Res. Lett. 2013, 40, 3649–3654. [CrossRef] 41. Hall, D.K.; Box, J.E.; Casey, K.A.; Hook, S.J.; Shuman, C.A.; Steffen, K. Comparison of satellite-derived and in-situ observations of ice and snow surface temperatures over Greenland. Remote Sens. Environ. 2008, 112, 3739–3749. [CrossRef] 42. Fausto, R.S.; van As, D.; Mankoff, K.D.; Vandecrux, B.; Citterio, M.; Ahlstrom, A.P.; Andersen, S.B.; Colgan, W.; Karlsson, N.B.; Kjeldsen, K.K.; et al. Programme for Monitoring of the Greenland Ice Sheet (PROMICE) automatic weather station data. Earth Syst. Sci. Data 2021, 13, 3819–3845. [CrossRef] 43. Ryan, J.; Smith, L.; Van As, D.; Cooley, S.; Cooper, M.; Pitcher, L.; Hubbard, A. Greenland Ice Sheet surface melt amplified by snowline migration and bare ice exposure. Sci. Adv. 2019, 5, eaav3738. [CrossRef] [PubMed] 44. Barnes, W.L.; Pagano, T.S.; Salomonson, V.V. Prelaunch characteristics of the Moderate Resolution Imaging Spectroradiometer (MODIS) on EOS-AM1. IEEE Trans. Geosci. Remote Sens. 1998, 36, 1088–1100. [CrossRef] 45. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Key, J.R.; Koenig, L.S. A Satellite-Derived Climate-Quality Data Record of the Clear-Sky Surface Temperature of the Greenland Ice Sheet. J. Clim. 2012, 25, 4785–4798. [CrossRef] 46. Key, J.; Haefliger, M. Arctic ice surface temperature retrieval from AVHRR thermal channels. J. Geophys. Res. Atmos. 1992, 97, 5885–5893. [CrossRef] 47. Key, J.R.; Collins, J.B.; Fowler, C.; Stone, R.S. High-latitude surface temperature estimates from thermal satellite data. Remote Sens. Environ. 1997, 61, 302–309. [CrossRef] 48. Hall, D.K.; Williams, R.S., Jr.; Luthcke, S.B.; Digirolamo, N.E. Greenland ice sheet surface temperature, melt and mass loss: 2000–2006. J. Glaciol. 2008, 54, 81–93. [CrossRef] 49. Stroeve, J.; Steffen, K. Variability of AVHRR-derived clear-sky surface temperature over the Greenland ice sheet. J. Appl. Meteorol. 1998, 37, 23–31. [CrossRef] 50. Tedesco, M.; Fettweis, X.; van den Broeke, M.R.; van de Wal, R.S.W.; Smeets, C.J.P.P.; van de Berg, W.J.; Serreze, M.C.; Box, J.E. The role of albedo and accumulation in the 2010 melting record in Greenland. Environ. Res. Lett. 2011, 6, 014005. [CrossRef] 51. Gałas, ´ A. Impact of volcanic eruptions on the environment and climatic conditions in the area of Poland (Central Europe). Earth-Sci. Rev. 2016, 162, 58–64. [CrossRef] 52. Abdalati, W.; Steffen, K. Greenland ice sheet melt extent: 1979–1999. J. Geophys. Res. -Atmos. 2001, 106, 33983–33988. [CrossRef] 53. Neff, W.; Compo, G.P.; Ralph, F.M.; Shupe, M.D. Continental heat anomalies and the extreme melting of the Greenland ice surface in 2012 and 1889. J. Geophys. Res. -Atmos. 2014, 119, 6520–6536. [CrossRef] 54. Bennartz, R.; Shupe, M.D.; Turner, D.D.; Walden, V.P.; Steffen, K.; Cox, C.J.; Kulie, M.S.; Miller, N.B.; Pettersen, C. July 2012 Greenland melt extent enhanced by low-level liquid clouds. Nature 2013, 496, 83–86. [CrossRef] [PubMed] 55. Cullather, R.I.; Andrews, L.C.; Croteau, M.J.; Digirolamo, N.E.; Hall, D.K.; Lim, Y.-K.; Loomis, B.D.; Shuman, C.A.; Nowicki, S.M.J. Anomalous Circulation in July 2019 Resulting in Mass Loss on the Greenland Ice Sheet. Geophys. Res. Lett. 2020, 47, e2020GL087263. [CrossRef] 56. Tedesco, M.; Fettweis, X. Unprecedented atmospheric conditions (1948–2019) drive the 2019 exceptional melting season over the Greenland ice sheet. Cryosphere 2020, 14, 1209–1223. [CrossRef] 57. Hearty, T.J., III; Lee, J.N.; Wu, D.L.; Cullather, R.; Blaisdell, J.M.; Susskind, J.; Nowicki, S.M.J. Intercomparison of Surface Temperatures from AIRS, MERRA, and MERRA-2 with NOAA and GC-Net Weather Stations at Summit, Greenland. J. Appl. Meteorol. Climatol. 2018, 57, 1231–1245. [CrossRef] 58. Cullather, R.I.; Nowicki, S.M.J. Greenland Ice Sheet Surface Melt and Its Relation to Daily Atmospheric Conditions. J. Clim. 2018, 31, 1897–1919. [CrossRef] 59. Sasgen, I.; Wouters, B.; Gardner, A.S.; King, M.D.; Tedesco, M.; Landerer, F.W.; Dahle, C.; Save, H.; Fettweis, X. Return to rapid ice loss in Greenland and record loss in 2019 detected by the GRACE-FO satellites. Commun. Earth Environ. 2020, 1, 8. [CrossRef] 60. Kimball, J.S.; Du, J.; Meierbachtol, T.W.; Kim, Y.; Johnson, J.V. Comparing Greenland Ice Sheet Melt Variability from Different Satellite Passive Microwave Remote Sensing Products Over a Common 5-year Record. Front. Earth Sci. 2021, 9, 654220. [CrossRef] http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Land Multidisciplinary Digital Publishing Institute

Assessment of MODIS Surface Temperature Products of Greenland Ice Sheet Using In-Situ Measurements

Download PDF
14 pages

Loading next page...
 
/lp/multidisciplinary-digital-publishing-institute/assessment-of-modis-surface-temperature-products-of-greenland-ice-fYHMy00VOi

References (37)

Publisher
Multidisciplinary Digital Publishing Institute
Copyright
© 1996-2022 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer The statements, opinions and data contained in the journals are solely those of the individual authors and contributors and not of the publisher and the editor(s). MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Terms and Conditions Privacy Policy
ISSN
2073-445X
DOI
10.3390/land11050593
Publisher site
See Article on Publisher Site

Abstract

land Article Assessment of MODIS Surface Temperature Products of Greenland Ice Sheet Using In-Situ Measurements 1 , † 1 , † 2 1 1 1 Xiaoge Yu , Tingting Wang , Minghu Ding , Yetang Wang , Weijun Sun , Qinglin Zhang 1 , and Baojuan Huai * College of Geography and Environment, Shandong Normal University, Jinan 250061, China; 201914010518@stu.sdnu.edu.cn (X.Y.); 201914010317@stu.sdnu.edu.cn (T.W.); 615002@sdnu.edu.cn (Y.W.); 612033@sdnu.edu.cn (W.S.); zhangqinglin@stu.sdnu.edu.cn (Q.Z.) State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China; dingmh@cma.gov.cn * Correspondence: 616070@sdnu.edu.cn or huaibaojuan@126.com † These authors contributed equally to this work. Abstract: Satellite-based data have promoted the research progress in polar regions under global climate change, meanwhile the uncertainties and limitations of satellite-derived surface temperatures are widely discussed over Greenland. This study validated the accuracy of ice surface temperature (IST) from the moderate-resolution imaging spectroradiometer (MODIS) over the Greenland ice sheet (GrIS). Daily MODIS IST was validated against the observational surface temperature from 24 automatic weather stations (AWSs) using the mean bias (MB), the root mean square (RMSE), and the correlation coefficient (R). The temporal and spatial variability over the GrIS spanning from March 2000 to December 2019 and the IST melt threshold (1 C) were analyzed. Generally, the MODIS IST was underestimated by an average of 2.68 C compared to AWSs, with cold bias mainly occurring in winter. Spatially, the R and RMSE performed the better accuracy of MODIS IST on the northwest, northeast, and central part of the GrIS. Furthermore, the mean IST is mainly concentrated between 20 C and 10 C in summer while between 50 C and 30 C in winter. Citation: Yu, X.; Wang, T.; Ding, M.; Wang, Y.; Sun, W.; Zhang, Q.; Huai, B. The largest positive IST anomalies (exceeds 3 C) occurred in southwestern GrIS during 2010. IST Assessment of MODIS Surface shows the positive trends mainly in spring and summer and negative in autumn and winter. Temperature Products of Greenland Ice Sheet Using In-Situ Measurements. Keywords: MODIS; ice surface temperature (IST); validation; Greenland ice sheet (GrIS); variability Land 2022, 11, 593. https://doi.org/ 10.3390/land11050593 Academic Editor: Nir Krakauer 1. Introduction Received: 16 March 2022 6 2 The Greenland ice sheet (GrIS) covers an area of 1.7  10 km and is the second Accepted: 16 April 2022 largest ice sheet in the world after the Antarctic ice sheet [1]. The GrIS had been rela- Published: 19 April 2022 tively stable during the 1970s, 1980s, and the early 1990s with the accumulation and loss Publisher’s Note: MDPI stays neutral staying relatively balanced [2,3]. In recent decades, there has been much attention paid with regard to jurisdictional claims in to the GrIS due to its enhanced melting [4–6], particularly relevant to abnormal climatic published maps and institutional affil- conditions [7–9] and accelerated mass loss [10–13]. The enhanced ablation was measured iations. using infrared and passive-microwave records [14–18]. The analysis of in-situ observations, satellite data, and regional climate models (RCMs) all revealed the accelerating GrIS mass loss [10,19]. The GrIS mass balance has decreased more than twice since the 21st century comparing with the last century [20,21]. The mass loss of GrIS over the past decade is about Copyright: © 2022 by the authors. six times that of the 1980s and the GrIS has contributed 13.7  1.1 mm to the global sea level Licensee MDPI, Basel, Switzerland. rise since 1972 [2]. If the acceleration continues, several studies using numerical model This article is an open access article strategies to predict the future evolution of GrIS suggested that GrIS could contribute to a distributed under the terms and rise in global sea level of 2.3 cm by 2050 [22], and 3.5 cm to 76.4 cm by 2300 [23,24]. Surface conditions of the Creative Commons temperature is an important factor in controlling mass balance [25]. Recently, the increase Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ in GrIS mass loss is mainly because of the increased surface meltwater runoff [26], and pro- 4.0/). longed surface temperature elevate exacerbates melting [27]. Surface temperature affects Land 2022, 11, 593. https://doi.org/10.3390/land11050593 https://www.mdpi.com/journal/land Land 2022, 11, 593 2 of 14 the basal melt [28] as its rise is an important step in the melt process [29]. Thus, tracking IST can better account for surface melt, mass balance as well as ice sheet surface processes. Currently, observational stations, regional climate models, and remote sensing are the three main sources to obtain IST for studying [30–33]. The AWSs on the ice sheet provide observations of the GrIS, for example, the Greenland Climate Network (GC-Net) [1,34] and the Greenland Ice Sheet Monitoring Program (PROMICE), which monitors surface and air temperature [35–37]. Meanwhile, existing coastal and low-density weather stations on the ice sheet make it hard to get accurate surface temperature observations of the entire GrIS. The difficulties of manipulating station devices in severe polar regions also make it hard to obtain continuous and complete observations record. RCMs can compensate for data recorded at certain periods and locations which could not be obtained from weather stations, especially in the polar regions [38]. RCMs are reliable in the entire GrIS context, however, in the regional and local context, climate variability can be missed occasionally [39,40]. Hence, the way to measure the surface temperature through satellite remote sensing is more outstanding than AWSs and RCMs with high temporal and spatial resolution over the GrIS. Nonetheless, the most practical way to get a wide range of surface temperature data is through the satellite remote sensing [41]. Accurate determination of satellite-derived IST will improve modeling of ice sheet melt and other ice sheet processes and changes. The GrIS is sensitive to climate change and plays a vital role in the global climate [1]. Hall showed an increase IST of ~0.55  0.44 C/decade and the northeastern GrIS experi- enced the largest increase of ~0.95  0.44 C/decade [8]. Under these conditions, detailed and accurate monitoring of the surface temperature on the entire GrIS is of significance. Recently, a multilayer, daily IST-albedo-water vapor product of Greenland, spanning from March 2000 to December 2019, has been developed. This paper mainly validated the accuracy of the MODIS IST products based on 24 AWSs from PROMICE. A description of the data and methods used in this study were given in 2. In 3, the IST performance was first demonstrated, and then based on its favorable performance, the temporal and spatial variability of GrIS IST, IST anomalies, and trends were further reproduced. Finally, in 4, the MODIS IST melt threshold 1 C using daily ablation data from 8 AWSs was also analyzed. 2. Materials and Methods 2.1. Observational Data The 24 AWSs (Figure 1) data used in this study are from the PROMICE, which is operated by the Geological Survey of Denmark and Greenland (GEUS) in collaboration with the National Space Institute at the Technical University of Denmark [35]. The time series of the AWS has different observation periods and completeness (Table 1). In this study, the daily IST data was derived from measured downward and upward longwave irradiance and surface emissivity is set to 0.97 [42]. Mostly located in the coastal area of the GrIS, the spatial layout of the AWSs is uneven. Roughly 25% of the stations are located below 500 m and only 16.7% are above 1000 m (Table 1). Moreover, only EGP, KAN_U, and CEN are located at the accumulation area according to the ELA (equilibrium line altitude) of the corresponding basin in 2017 [43]. Table 1. Geographical characteristics of the meteorological stations used in this study. Station Latitude Longitude Elevation Elevation-ELA Start Date Name ( N) ( W) (m) (m) KPC_L * 79.91 24.08 370 707 17 July 2008 KPC_U 79.83 25.17 870 207 17 July 2008 EGP 75.62 35.97 2660 1583 1 May 2016 SCO_L 72.22 26.82 460 1192 21 July 2008 SCO_U 72.39 27.23 970 682 21 July 2008 MIT 65.69 37.83 440 970 3 May 2009 TAS_L 65.64 38.90 250 1160 28 August 2007 Land 2022, 11, 593 3 of 14 Table 1. Cont. Station Latitude Longitude Elevation Elevation-ELA Start Date Name ( N) ( W) (m) (m) TAS_U 65.70 38.87 570 840 15 August 2007 TAS_A 65.78 38.90 890 520 23 August 2013 QAS_L 61.03 46.85 280 1190 24 August 2007 QAS_M 61.10 46.83 630 840 11 August 2016 QAS_U 61.18 46.82 900 570 7 August 2008 QAS_A 61.24 46.73 1000 470 20 August 2012 NUK_L 64.48 49.54 530 885 20 August 2007 NUK_U 64.51 49.27 1120 295 20 August 2007 NUK_N 64.95 49.89 920 495 25 July 2010 KAN_L 67.10 49.95 670 745 1 September 2008 KAN_M 67.07 48.84 1270 145 2 September 2008 KAN_U 67.00 47.03 1840 425 4 April 2009 UPE_L 72.89 54.30 220 843 17 August 2009 UPE_U 72.89 53.58 940 123 17 August 2009 THU_L 76.40 68.27 570 493 9 August 2010 THU_U 76.42 68.15 760 303 9 August 2010 Land 2021, 10, x FOR PEER REVIEW 3 of 16 CEN 77.17 61.11 1880 817 23 May 2017 * L: Lower station, M: Middle station, U: Upper station. Figure 1. Map of Greenland showing the elevation and locations of the PROMICE AWS. Six tran- Figure 1. Map of Greenland showing the elevation and locations of the PROMICE AWS. Six transects sects are magnified in subplots. are magnified in subplots. 2.2. MODIS IST Products Table 1. Geographical characteristics of the meteorological stations used in this study. MODIS is a polar-orbiting, 36-channel, and across-track scanning spectro-radiometer Station Elevation- whose images cover Latituthe de (whole ° N) Longi region tudof e (the ° W)planet Elevatevery ion (m) one to two days [44 S]. taMODIS rt Date has Name ELA (m) been flying aboard NASA Earth Observation System (EOS) Terra satellites since 2000 and KPC_L * 79.91 24.08 370 −707 17 July 2008 Aqua since 2002 [45], so the GrIS swath-based daily gridded products began to be available KPC_U 79.83 25.17 870 −207 17 July 2008 at the beginning of 2000. Terra’s orbit around the Earth is timed so that it passes across EGP 75.62 35.97 2660 1583 1 May 2016 Greenland in the afternoon (14:30–16:30 UTC) [41]. MODIS Collection 6.1 IST used in this SCO_L 72.22 26.82 460 −1192 21 July 2008 study is from multilayer, daily IST-albedo-water vapor products developed using standard MODIS SCO_ datasets U from 72.3the 9 Terra satellite 27.23 [28]. 970 −682 21 July 2008 MIT 65.69 37.83 440 −970 3 May 2009 TAS_L 65.64 38.90 250 −1160 28 August 2007 TAS_U 65.70 38.87 570 −840 15 August 2007 TAS_A 65.78 38.90 890 −520 23 August 2013 QAS_L 61.03 46.85 280 −1190 24 August 2007 QAS_M 61.10 46.83 630 −840 11 August 2016 QAS_U 61.18 46.82 900 −570 7 August 2008 QAS_A 61.24 46.73 1000 −470 20 August 2012 NUK_L 64.48 49.54 530 −885 20 August 2007 NUK_U 64.51 49.27 1120 −295 20 August 2007 NUK_N 64.95 49.89 920 −495 25 July 2010 KAN_L 67.10 49.95 670 −745 1 September 2008 KAN_M 67.07 48.84 1270 −145 2 September 2008 KAN_U 67.00 47.03 1840 425 4 April 2009 UPE_L 72.89 54.30 220 −843 17 August 2009 UPE_U 72.89 53.58 940 −123 17 August 2009 THU_L 76.40 68.27 570 −493 9 August 2010 THU_U 76.42 68.15 760 −303 9 August 2010 CEN 77.17 61.11 1880 817 23 May 2017 * L: Lower station, M: Middle station, U: Upper station. Land 2022, 11, 593 4 of 14 Its algorithm was derived from the algorithm developed for the MODIS sea ice product MOD29 or MYD29 (MOD and MYD stand for Terra and Aqua products, respec- tively) [45–47]. The “MOD29” IST products during March 2000~December 2019, provide daily and monthly mean ISTs with polar-stereographic grids at 0.78  0.78 km resolu- tion [28]. As the input product of the MOD29 algorithm, MOD35 determines the cloud obscuration of the MODIS IST products. The cloudy grid cells without IST value are excluded calculation process. 2.3. MODIS Products Preprocessing The cloud coverage affects the accuracy of MODIS IST products. The internal cloud cover incorrectly identified instances of the ice surface as cloudless on a lot of days in JJA (June, July, and August) [8]. Therefore, the ISTs are actually the cloud top temperatures, far below the clear surface temperature (up to 30 C) [8]. Therefore, we checked the MODSI IST of each grid over two standard deviations to reduce outliers caused by cloud masks or other reasons. The percentage of the removed pixels was shown in Table 2. Table 2. The percentage of the MODIS IST pixels that had been removed. Year Percentage (%) Year Percentage (%) 2000 0.56 2010 0.46 2001 0.55 2011 0.49 2002 0.52 2012 0.48 2003 0.52 2013 0.50 2004 0.52 2014 0.50 2005 0.49 2015 0.50 2006 0.48 2016 0.47 2007 0.50 2017 0.48 2008 0.50 2018 0.48 2009 0.48 2019 0.45 As the ice/snow surface cannot exist in the form of ice/snow above 0 C [8], the fake pixels values (0 C) found on the ice sheet are removed. The most likely reason for a few pixels with the wrong IST is incorrect cloud masking caused by MODIS cloud mask [48]. Additionally, some pixel values representing no data, cloud cover, and fill value, are also removed. 2.4. Statistical Indexes for MODIS IST Assessment The daily MODIS ISTs used to make a comparison with the observations were extracted respectively from the nearest gridded MODIS IST at the ice sheet. Meanwhile, three classical statistical indexes were used to estimate the performance of the MODIS product, including the bias, root mean square (RMSE), and the correlation coefficient (R). The calculation formulas are as follows: (M O ) i i i=1 Bias = (1) (O M ) i i i=1 RMSE = (2) å (O O)(M M) i i i=1 s s r = (3) N N 2 2 (O O) (M M) å å i i i=1 i=1 Land 2022, 11, 593 5 of 14 where N is the available number of the sample, O is the observed IST, and M is the corresponding MODIS IST. 2.5. Anomaly and Trend Analysis As the MODIS IST dataset of the GrIS used in this study is available since March 2000, the annual anomaly of 2000 was not calculated. For each grid, the annual IST anomalies were calculated using the monthly data, relative to the multi-year average annual IST from 2001 to 2019. The summer IST includes the year 2000. When seasonal and annual IST trends of each grid were calculated, only the grids with 20-year data are used to calculate. 3. Results 3.1. MODIS IST Performance To initially obtain the performance of MODIS IST, 12 AWSs are selected according to their location and relative data volume. Figure 2 shows that the MODIS IST agrees well with the observations, with the R > 0.84 (TAS_ L and QAS_L). A significant characteristic Land 2021, 10, x FOR PEER REVIEW 6 of 16 is MODIS IST has a cold deviation of 2–3 C. MODIS IST shows the best agreement with KPC_U AWS observation, with a low RMSE (RMSE = 3.71 C) and a high correlation (R = 0.97), whereas, generally poorer agreement with TAS_L and QAS_L. Figure 2. MODIS and observed daily IST, the red diagonal represents the 1:1 line. Statistics show Figure 2. MODIS and observed daily IST, the red diagonal represents the 1:1 line. Statistics show Correlation coefficient (R), root mean squared error (RMSE), and bias. Correlation coefficient (R), root mean squared error (RMSE), and bias. Figure 3 also suggests that the bias indicates the IST tends to underestimate surface Figure 3 also suggests that the bias indicates the IST tends to underestimate surface temperature (2–4 C). The R and RMSE indicate that MODIS IST is highly consistent with temperature (2–4 °C). The R and RMSE indicate that MODIS IST is highly consistent ground measurements. Spatially, the performance of MODIS IST in the north is better with ground measurements. Spatially, the performance of MODIS IST in the north is bet- than that in the south, with a higher average R (0.96) and lower average RMSE (4.19 C). ter than that in the south, with a higher average R (0.96) and lower average RMSE (4.19 Compared with other regions, the R and RMSE show a better accuracy of MODIS IST on °C). Compared with other regions, the R and RMSE show a better accuracy of MODIS the northwest, northeast, and central GrIS. In these areas, the R is close to 1 and the range IST on the northwest, northeast, and central GrIS. In these areas, the R is close to 1 and of RMSE is concentrated at 3.7–4.9 C, indicating a high correlation. The MODIS IST shows the range of RMSE is concentrated at 3.7–4.9 °C, indicating a high correlation. The the best agreement with observations in the northeast, with the highest average R (0.96) MODIS IST shows the best agreement with observations in the northeast, with the high- est average R (0.96) and the lowest average RMSE (3.93 °C), however, generally poorer agreement with AWSs observations in the southeast. Land 2022, 11, 593 6 of 14 Land 2021, 10, x FOR PEER REVIEW 7 of 16 and the lowest average RMSE (3.93 C), however, generally poorer agreement with AWSs observations in the southeast. Fig Figure ure 33. . Cor Corr relelation ation coeffi coefc fi ien cient t (R), (R), bibias, as, an and d roo root t m mean ean squa squar red ed er err ror or (RMS (RMSE) E) bet between ween da daily ily MODIS MODIS IST and observations from all AWSs (AWSs shown in bold are above the ELA). IST and observations from all AWSs (AWSs shown in bold are above the ELA). Figure 4 compares MODIS IST time series with AWSs to analyze the typical tem- Figure 4 compares MODIS IST time series with AWSs to analyze the typical temporal poral variation and (in) consistency of time series derived from MODIS IST. KPC_U, variation and (in) consistency of time series derived from MODIS IST. KPC_U, KAN_U, KAN_U, and EGP were selected for their high-continuity data. The MODIS IST perfor- and EGP were selected for their high-continuity data. The MODIS IST performance is good, mance is good, and the time series are consistent well with observations, only with a and the time series are consistent well with observations, only with a slight underestimation slight underestimation in summer. Notably, the cold bias of MODIS IST mainly occurs in Land 2021, 10, x FOR PEER REVIEW 8 of 16 in summer. Notably, the cold bias of MODIS IST mainly occurs in winter. The average cold winter. The average cold deviations in summer and winter are 2.04 °C and 4.35 °C, re- deviations in summer and winter are 2.04 C and 4.35 C, respectively. spectively. Figure 4. Time series of MODIS IST (red line) and surface temperature (green line) from EGP (a), Figure 4. Time series of MODIS IST (red line) and surface temperature (green line) from EGP (a), KAN_U KAN_U ((b b )), , aand nd KKPC_U PC_U (c). (c ). Overall, the performance of MODIS products is good, and the range of the cold bias is concentrated at 2–4 °C. A cold deviation of 0.98 °C was found when MODIS IST was compared with the observation data measured by Summit Station of Greenland [28]. The results of thermochron data on the Summit also showed a cold deviation of 3.14 °C of MODIS IST [45]. Given the limited coverage of AWSs on GrIS, we suggest that ISTs ob- tained from the MODIS IST products are characterized by high accuracy and could be reliably used as an alternative or supplement to Greenland temperature monitoring, es- pecially for IST in summer. Generally, fine space-based climate information obtained by analyzing MODIS IST can effectively evaluate the temporal and spatial variability of the GrIS and be applied in many fields. 3.2. Temporal and Spatial Variability of GrIS IST Figure 5 are clear-sky IST maps illustrating the MODIS IST on the GrIS at seasonal and annual scales. The MODIS IST revealed distinct spatial gradients varying from outer (low elevation) to inner (high elevation), showing a colder inside with higher elevation. Apparently, the IST in the northern GrIS is much lower than that in the southern region while there is little difference between the eastern and western GrIS. The mean autumn IST is a little colder than the spring and annual mainly reflected in the inner of the GrIS. The mean summer IST is mainly concentrated between −20 °C and −10 °C with the high- est IST (−1.33 °C) in the southwest margin of the GrIS and the lowest IST (−17.98 °C) in the inner. Conversely, the mean winter IST is mainly concentrated between −50 °C and −30 °C with the highest and lowest IST values of −9.85 °C and −50.13 °C, respectively. The minimum IST values in winter are found at high elevations just as in summer but Land 2022, 11, 593 7 of 14 Overall, the performance of MODIS products is good, and the range of the cold bias is concentrated at 2–4 C. A cold deviation of 0.98 C was found when MODIS IST was compared with the observation data measured by Summit Station of Greenland [28]. The results of thermochron data on the Summit also showed a cold deviation of 3.14 C of MODIS IST [45]. Given the limited coverage of AWSs on GrIS, we suggest that ISTs obtained from the MODIS IST products are characterized by high accuracy and could be reliably used as an alternative or supplement to Greenland temperature monitoring, especially for IST in summer. Generally, fine space-based climate information obtained by analyzing MODIS IST can effectively evaluate the temporal and spatial variability of the GrIS and be applied in many fields. 3.2. Temporal and Spatial Variability of GrIS IST Figure 5 are clear-sky IST maps illustrating the MODIS IST on the GrIS at seasonal and annual scales. The MODIS IST revealed distinct spatial gradients varying from outer (low elevation) to inner (high elevation), showing a colder inside with higher elevation. Apparently, the IST in the northern GrIS is much lower than that in the southern region while there is little difference between the eastern and western GrIS. The mean autumn IST is a little colder than the spring and annual mainly reflected in the inner of the GrIS. The mean summer IST is mainly concentrated between 20 C and 10 C with the highest IST (1.33 C) in the southwest margin of the GrIS and the lowest IST (17.98 C) Land 2021, 10, x FOR PEER REVIEW 9 of 16 in the inner. Conversely, the mean winter IST is mainly concentrated between 50 C and 30 C with the highest and lowest IST values of 9.85 C and 50.13 C, respectively. The minimum IST values in winter are found at high elevations just as in summer but tend to migrate farther north, indicating that the coldest ISTs are not always located at tend to migrate farther north, indicating that the coldest ISTs are not always located at the the highest elevations, consistent with the previous study [49]. highest elevations, consistent with the previous study [49]. Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter (d) (d) and the mean annual MODIS IST (e) from March 2000 to December 2019. and the mean annual MODIS IST (e) from March 2000 to December 2019. 3.3. GrIS IST Anomalies 3.3. GrIS IST Anomalies Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 re- respectively. The IST anomalies between 1 C and 1 C are dominated in most of the spectively. The IST anomalies between −1 °C and 1 °C are dominated in most of the study period except for some abnormally warm and cold years. The largest positive IST study period except for some abnormally warm and cold years. The largest positive IST anomalies (exceeds 3 C) occurred in the southwest in 2010 when almost the entire GrIS anomalies (exceeds 3 °C) occurred in the southwest in 2010 when almost the entire GrIS experienced positive IST anomalies consistent with [8]. In that year, Greenland experienced experienced positive IST anomalies consistent with [8]. In that year, Greenland experi- melting up to another 60 days compared with the 1960–2010 average (1980–2010 were enced melting up to another 60 days compared with the 1960–2010 average (1980–2010 simulated data), with the largest differences occurring at the southwestern and western were simulated data), with the largest differences occurring at the southwestern and margins of the GrIS [6]. The previous study shows that the unusual melt event in 2010 was western margins of the GrIS [6]. The previous study shows that the unusual melt event trigged by large positive air temperature anomalies during May, accelerating snowpack in 2010 was trigged by large positive air temperature anomalies during May, accelerat- metamorphism, and premature bare ice exposure [50]. The GrIS IST anomaly in May 2010 ing snowpack metamorphism, and premature bare ice exposure [50]. The GrIS IST may be relevant to the eruption of Eyjafjallajokull ˝ volcano on March 2010 which once had a anomaly in May 2010 may be relevant to the eruption of Eyjafjallajőkull volcano on profound influence on the environment and climatic conditions of neighboring regions [51]. March 2010 which once had a profound influence on the environment and climatic con- With the positive albedo feedback and the fact that the wet snow could absorb up to three ditions of neighboring regions [51]. With the positive albedo feedback and the fact that times more incident solar energy than dry snow, the further melt was fostered [52]. the wet snow could absorb up to three times more incident solar energy than dry snow, the further melt was fostered [52]. Figure 6. Annual IST anomalies (°C) calculated from monthly-derived MODIS IST. In addition to 2010, relatively large positive IST anomalies (1–3 °C) were mainly found in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found in central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire GrIS in 2016 while northern and western GrIS in 2019. By comparison, a relatively large Land 2021, 10, x FOR PEER REVIEW 9 of 16 tend to migrate farther north, indicating that the coldest ISTs are not always located at the highest elevations, consistent with the previous study [49]. Figure 5. Seasonal MODIS IST over the GrIS in spring (a), summer (b), autumn (c), and winter (d) and the mean annual MODIS IST (e) from March 2000 to December 2019. 3.3. GrIS IST Anomalies Figure 6 shows the temporal and spatial annual IST anomaly from 2001 to 2019 re- spectively. The IST anomalies between −1 °C and 1 °C are dominated in most of the study period except for some abnormally warm and cold years. The largest positive IST anomalies (exceeds 3 °C) occurred in the southwest in 2010 when almost the entire GrIS experienced positive IST anomalies consistent with [8]. In that year, Greenland experi- enced melting up to another 60 days compared with the 1960–2010 average (1980–2010 were simulated data), with the largest differences occurring at the southwestern and western margins of the GrIS [6]. The previous study shows that the unusual melt event in 2010 was trigged by large positive air temperature anomalies during May, accelerat- ing snowpack metamorphism, and premature bare ice exposure [50]. The GrIS IST anomaly in May 2010 may be relevant to the eruption of Eyjafjallajőkull volcano on March 2010 which once had a profound influence on the environment and climatic con- ditions of neighboring regions [51]. With the positive albedo feedback and the fact that Land 2022, 11, 593 the wet snow could absorb up to three times more incident solar energy than d 8 of ry 14 snow, the further melt was fostered [52]. Land 2021, 10, x FOR PEER REVIEW 10 of 16 Figure 6. Annual IST anomalies ( C) calculated from monthly-derived MODIS IST. Figure 6. Annual IST anomalies (°C) calculated from monthly-derived MODIS IST. In addition to 2010, relatively large positive IST anomalies (1–3 C) were mainly found In addition to 2010, relatively large positive IST anomalies (1–3 °C) were mainly region of negative IST anomalies that between –1 °C and –3 °C were found in 2001, 2011, in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found in found in 2003, 2012, 2016, and 2019. The high positive IST anomalies were mainly found central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire GrIS in 2013 and 2015. Especially, the largest negative IST anomalies (below −3 °C) was found in in central and southern GrIS in 2003, southeastern GrIS in 2012, and almost the entire 2016 while northern and western GrIS in 2019. By comparison, a relatively large region of 2015 in southwestern GrIS. GrIS negative in 201IST 6 wh anomalies ile north that ern between and western –1 C G and rIS–3 in C 20wer 19. eBfound y comin pa 2001, rison 2011, , a re2013 latively and large Figure 7 shows the temporal and spatial characteristics of the mean summer IST 2015. Especially, the largest negative IST anomalies (below 3 C) was found in 2015 in anomaly from 2000 to 2019, respectively. As reported by Hall et al. [8], the summer of southwestern GrIS. 2012 is the warmest summer, with the southeast experiencing the largest positive IST Figure 7 shows the temporal and spatial characteristics of the mean summer IST anom anomaly alies (ex from ceed 2000 ing to 3 2019, °C)r,espectively related to . As th re eported strongest by Hall No etrth al. [A 8],m the erica summer n hea of t 2012 wave since is the warmest summer, with the southeast experiencing the largest positive IST anomalies 1895 [53]. In 2012, almost the entire GrIS (98.6%) experienced an extreme melting event., (exceeding 3 C), related to the strongest North American heat wave since 1895 [53]. In 2012, even at Summit [7]. This extreme melt event was associated with an abnormal ridge of almost the entire GrIS (98.6%) experienced an extreme melting event., even at Summit [7]. warm air that was stagnant over Greenland and the radiative effects of low-level liquid This extreme melt event was associated with an abnormal ridge of warm air that was layer clouds [7,54]. Previous studies had shown that on 11–13 July, the GrIS experienced stagnant over Greenland and the radiative effects of low-level liquid layer clouds [7,54]. the most extensive melt extent based on satellite-derived data [7,9]. Figure 8 are the Previous studies had shown that on 11–13 July, the GrIS experienced the most extensive MOD melt IS extent IST on based 11–1 on 3 Jsatellite-derived uly of the GrIS data sho [7 wi ,9]. ng Figur thae t 8 th ar ee IS the T MODIS is conce IST ntra onted 11–13 beJuly tween 0 °C of the GrIS showing that the IST is concentrated between 0 C and 3 C, especially the and −3 °C, especially the western GrIS experienced the largest area of highest IST. western GrIS experienced the largest area of highest IST. Figure 7. Summer IST anomalies ( C) calculated from monthly IST (2000 to 2019). Figure 7. Summer IST anomalies (°C) calculated from monthly IST (2000 to 2019). Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is under cloud cover or has no data is considered as no data. In addition, a relatively large area of positive IST anomalies (0–2 °C) in summer was mainly found in 2007, 2010, 2016 and 2019. The recent warm summer found in our study period is the summer of 2019, agreeing with the previous study [55,56]. The increasing Land 2021, 10, x FOR PEER REVIEW 10 of 16 region of negative IST anomalies that between –1 °C and –3 °C were found in 2001, 2011, 2013 and 2015. Especially, the largest negative IST anomalies (below −3 °C) was found in 2015 in southwestern GrIS. Figure 7 shows the temporal and spatial characteristics of the mean summer IST anomaly from 2000 to 2019, respectively. As reported by Hall et al. [8], the summer of 2012 is the warmest summer, with the southeast experiencing the largest positive IST anomalies (exceeding 3 °C), related to the strongest North American heat wave since 1895 [53]. In 2012, almost the entire GrIS (98.6%) experienced an extreme melting event., even at Summit [7]. This extreme melt event was associated with an abnormal ridge of warm air that was stagnant over Greenland and the radiative effects of low-level liquid layer clouds [7,54]. Previous studies had shown that on 11–13 July, the GrIS experienced the most extensive melt extent based on satellite-derived data [7,9]. Figure 8 are the MODIS IST on 11–13 July of the GrIS showing that the IST is concentrated between 0 °C and −3 °C, especially the western GrIS experienced the largest area of highest IST. Land 2022, 11, 593 9 of 14 In addition, a relatively large area of positive IST anomalies (0–2 C) in summer was mainly found in 2007, 2010, 2016 and 2019. The recent warm summer found in our study period is the summer of 2019, agreeing with the previous study [55,56]. The increasing IST is the basic factor causing the increasing GrIS melting [57]. The warm, moist air intrusions could potentially drive melt events [54,58]. At the end of July 2019, the GrIS melting extent achieved more than 62%, even at Summit [55] and the ablation zones continued through August and September, entering accumulation conditions in early October [59]. Contrary to the above-average warmth conditions during the summers mentioned above, relatively Figure 7. Summer IST anomalies (°C) calculated from monthly IST (2000 to 2019). large areas of negative IST anomalies were found in 2000, 2001, 2002, 2006, 2009, 2013, 2017 and 2018. Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is Figure 8. MODIS IST maps for (a): 11 July, (b): 12 July, and (c): 13 July 2012. The IST grid that is under cloud cover or has no data is considered as no data. under cloud cover or has no data is considered as no data. 3.4. IST Trend Analysis of GrIS Figure 9 shows the positive IST trends mainly in spring and summer and negative In addition, a relatively large area of positive IST anomalies (0–2 °C) in summer was trends in fall and winter. The annual IST trends are much more moderate than the seasonal mainly found in 200IST 7, 2 tr0 ends 10, and 201the 6 amost nd 2IST 019 tr.ends The arre e concentrated cent warm ~+ sum 0.5mC/decade, er found showing in our ast lowe udy r spatial variability. On an annual scale, the northern GrIS mainly experienced positive IST period is the summer of 2019, agreeing with the previous study [55,56]. The increasing trends (0~0.5 C/decade) while the southern GrIS mainly experienced negative IST trends (1~0 C/decade). Autumn cooling was common and Hall et al. [8] also found a decrease in IST (~1.49  1.20 C/decade) in autumn during the period 2000–2012. Similarly, this work also showed the northwestern GrIS experienced positive IST trends except autumn during the study period, especially in spring with the positive IST trends reaching up to 1.5 C/decade. Furthermore, the minimum negative IST trends (below 1.5 C/decade) were found in the small area of southern and eastern GrIS in winter and a small area of southern GrIS in autumn. Land 2021, 10, x FOR PEER REVIEW 11 of 16 IST is the basic factor causing the increasing GrIS melting.[57]. The warm, moist air in- trusions could potentially drive melt events [54,58]. At the end of July 2019, the GrIS melting extent achieved more than 62%, even at Summit [55] and the ablation zones con- tinued through August and September, entering accumulation conditions in early Octo- ber [59]. Contrary to the above-average warmth conditions during the summers men- tioned above, relatively large areas of negative IST anomalies were found in 2000, 2001, 2002, 2006, 2009, 2013, 2017 and 2018. 3.4. IST Trend Analysis of GrIS Figure 9 shows the positive IST trends mainly in spring and summer and negative trends in fall and winter. The annual IST trends are much more moderate than the sea- sonal IST trends and the most IST trends are concentrated ~+−0.5 °C/decade, showing a lower spatial variability. On an annual scale, the northern GrIS mainly experienced posi- tive IST trends (0~0.5 °C/decade) while the southern GrIS mainly experienced negative IST trends (−1~0 °C/decade). Autumn cooling was common and Hall et al. [8] also found a decrease in IST (~−1.49 ± 1.20 °C/decade) in autumn during the period 2000–2012. Simi- larly, this work also showed the northwestern GrIS experienced positive IST trends ex- cept autumn during the study period, especially in spring with the positive IST trends reaching up to 1.5 °C/decade. Furthermore, the minimum negative IST trends (below −1.5 °C/decade) were found in the small area of southern and eastern GrIS in winter and Land 2022, 11, 593 10 of 14 a small area of southern GrIS in autumn. Figure 9. Seasonal MODIS IST trends over the GrIS in spring (a), summer (b), autumn (c) and winter Figure 9. Seasonal MODIS IST trends over the GrIS in spring (a), summer (b), autumn (c) and win- (d) from March 2000 to December 2019 and the annual MODIS IST (e). ter (d) from March 2000 to December 2019 and the annual MODIS IST (e). 4. Discussion 4. Discussion GrIS melt and runoff have increased rapidly since the early 1990s [26]. The low density GrIS melt and runoff have increased rapidly since the early 1990s [26]. The low and discontinuity of AWS measurement are the inherent limitation in evaluating the melt dens onity the aentir nd deisice con sheet. tinuity Surface of AWS melting measurement conditionsare that thcannot e inherbe ent measur limitaed tiodue n in to evthe aluating AWSs constraints can only be obtained by remote sensing [60]. Given the significance of the melt on the entire ice sheet. Surface melting conditions that cannot be measured due remotely sensed melt products for monitoring GrIS melt, an assessment of the MODIS to the AWSs constraints can only be obtained by remote sensing [60]. Given the signifi- IST-derived melt data is essential. cance of remotely sensed melt products for monitoring GrIS melt, an assessment of the The MODIS IST-derived melt data is determined by the melt threshold (1 C) and MODIS IST-derived melt data is essential. a non-cloud-obscured IST grid that is 1 C is defined as “melt”. Therefore, the melt The MODIS IST-derived melt data is determined by the melt threshold (−1 °C) and threshold is the key factor to be validated. The IST products have an accuracy of 1 C a non-cloud-obscured IST grid that is ≥−1 °C is defined as “melt”. Therefore, the melt and melt may occur when temperatures are slightly below freezing with a strong solar threshold is the key factor to be validated. The IST products have an accuracy of ±1 °C radiation [28]. Given the completeness and amount of the AWSs daily ablation data, the daily ablation data from 5 AWSs at low elevations and 3 AWSs at high elevations were and melt may occur when temperatures are slightly below freezing with a strong solar selected to validate the melt threshold. The daily ablation data was set to 25 mm based radiation [28]. Given the completeness and amount of the AWSs daily ablation data, the on the accuracy of the pressure transducer [37]. Only the data that were both available daily ablation data from 5 AWSs at low elevations and 3 AWSs at high elevations were for MODIS and AWS were used in this study. The days that ware detected melt or no melt both from MODIS and AWS were recorded as “BB” and these days can be considered well-matched days. If the days were detected melt only from MODIS or AWS, then they were recorded as “MM” or “SM” and these days can be considered failed-matched days. Figure 10 shows the day-to-day melt agreement conditions between MODIS and AWSs. Obviously, AWSs at low elevations reproduced more melt occasions than at high elevations. In the strong ablation year of 2012, except for the KPC_L, MODIS IST reproduced fewer melt conditions than AWSs. In general, MODIS IST performed an average agreement of 55% with the highest agreement in KPC_L (61%), located at the northeastern GrIS. NUK_L at southwestern GrIS had the lowest consistency of 41% with much fewer melt days than that detected by MODIS IST. Other AWSs show an agreement between 50% and 60%. Overall, the 1 C melt threshold suggested a 41–61% consistency of melt conditions (BB) between MODIS and AWSs, and generally, fewer MODIS melt days occurred than AWSs. Given the fewer MODIS melt days based on 1 C and the cold bias found in this work, lower melt thresholds were used to possibly achieve higher consistencies (Figure 11). The results showed that consistencies were 46–66% and 51–70% based on melt thresholds respectively 1.5 C and 2 C, respectively. With the melt threshold decreasing, the consistency was generally becoming higher especially for the stations QAS_L, NUK_L, KAN_L, and TAS_U while KPC_L and QAS_U had lower consistencies. Land 2021, 10, x FOR PEER REVIEW 12 of 16 selected to validate the melt threshold. The daily ablation data was set to 25 mm based on the accuracy of the pressure transducer [37]. Only the data that were both available for MODIS and AWS were used in this study. The days that ware detected melt or no melt both from MODIS and AWS were recorded as “BB” and these days can be consid- ered well-matched days. If the days were detected melt only from MODIS or AWS, then they were recorded as “MM” or “SM” and these days can be considered failed-matched days. Figure 10 shows the day-to-day melt agreement conditions between MODIS and AWSs. Obviously, AWSs at low elevations reproduced more melt occasions than at high elevations. In the strong ablation year of 2012, except for the KPC_L, MODIS IST repro- duced fewer melt conditions than AWSs. In general, MODIS IST performed an average agreement of 55% with the highest agreement in KPC_L (61%), located at the northeast- ern GrIS. NUK_L at southwestern GrIS had the lowest consistency of 41% with much fewer melt days than that detected by MODIS IST. Other AWSs show an agreement be- tween 50% and 60%. Overall, the −1 °C melt threshold suggested a 41–61% consistency of melt conditions (BB) between MODIS and AWSs, and generally, fewer MODIS melt days occurred than AWSs. Given the fewer MODIS melt days based on −1 °C and the cold bias found in this work, lower melt thresholds were used to possibly achieve higher consistencies (Figure 11). The results showed that consistencies were 46–66% and 51– 70% based on melt thresholds respectively −1.5 °C and −2 °C, respectively. With the melt threshold decreasing, the consistency was generally becoming higher especially for the Land 2022, 11, 593 11 of 14 stations QAS_L, NUK_L, KAN_L, and TAS_U while KPC_L and QAS_U had lower con- sistencies. Land 2021, 10, x FOR PEER REVIEW 13 of 16 Fig Figure ure 1 10. 0. Th The e c consistency onsistency b between etween MODI MODIS S a and nd A AWSs WSs melt melt condition. condition. Figure Figure 11. 11. Consistency Consistency of of melt melt conditions conditions (BB) (BB) between between MODIS MODIS and and station station based based on on the the melt melt threshold −1 °C, −1.5 °C, and −2 °C. threshold 1 C, 1.5 C, and 2 C. 5. Conclusions 5. Conclusions In this study, based on PROMICE AWSs observations, three statistical indexes, bias, In this study, based on PROMICE AWSs observations, three statistical indexes, bias, RMSE, and R are used to comprehensively evaluate the performance of the MODIS IST over RMSE, and R are used to comprehensively evaluate the performance of the MODIS IST the GrIS. Generally, MODIS IST has a widespread cold bias over the GrIS, ranging from 2 over the GrIS. Generally, MODIS IST has a widespread cold bias over the GrIS, ranging from −2 to −4 °C and this cold bias mainly occurred in winter. The consistency between MODIS IST and in-situ IST is very high on the GrIS, with an average of R = 0.92 and RMSE = 4.36 °C. Spatially, the R and RMSE agree on a better accuracy of MODIS IST on the northwestern, northeastern, and central GrIS. The MODIS IST can well demonstrate the temporal and spatial variability of GrIS. The mean summer IST is mainly concentrated between −20 °C and −10 °C with the high- est and lowest IST values of −1.33 °C and −17.98 °C, respectively. Besides, the largest positive IST anomalies (exceeding 3 °C) occurred in southwestern GrIS in 2010. Summer 2012 was the warmest summer with the largest positive IST anomalies (exceeding 3 °C) in the southeastern GrIS. The positive IST trends are mainly in spring and summer and the negative IST trends are in autumn and winter. Importantly, most of the annual IST trends are concentrated between ~±0.5 °C/decade. This work also verifies the rationality of the MODIS IST melt threshold of −1 °C us- ing the daily ablation data from AWSs. The results show that the −1 °C melt threshold suggested a 41–61% consistency of melt conditions between MODIS and AWSs. Fur- thermore, with the melt threshold decreasing (−1.5 °C, −2 °C), the consistency is general- ly becoming higher. Author Contributions: X.Y.: Formal analysis, visualization, writing—original draft. T.W.: formal analysis, visualization, writing—original draft. M.D.: formal analysis, writing—original draft. Y.W.: formal analysis, methodology. W.S.: formal analysis, methodology. Q.Z.: methodology, software, validation. B.H.: funding acquisition, supervision, conceptualization, methodology, val- idation, writing—review and editing. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Natural Science Foundation of China grant number 42171121, 41701059. And The APC was funded by 42171121. Data Availability Statement: All data and the script of the whole processes are available through an email request to the authors. Land 2022, 11, 593 12 of 14 to 4 C and this cold bias mainly occurred in winter. The consistency between MODIS IST and in-situ IST is very high on the GrIS, with an average of R = 0.92 and RMSE = 4.36 C. Spatially, the R and RMSE agree on a better accuracy of MODIS IST on the northwestern, northeastern, and central GrIS. The MODIS IST can well demonstrate the temporal and spatial variability of GrIS. The mean summer IST is mainly concentrated between 20 C and 10 C with the highest and lowest IST values of 1.33 C and 17.98 C, respectively. Besides, the largest positive IST anomalies (exceeding 3 C) occurred in southwestern GrIS in 2010. Summer 2012 was the warmest summer with the largest positive IST anomalies (exceeding 3 C) in the southeastern GrIS. The positive IST trends are mainly in spring and summer and the negative IST trends are in autumn and winter. Importantly, most of the annual IST trends are concentrated between ~0.5 C/decade. This work also verifies the rationality of the MODIS IST melt threshold of 1 C using the daily ablation data from AWSs. The results show that the 1 C melt threshold suggested a 41–61% consistency of melt conditions between MODIS and AWSs. Further- more, with the melt threshold decreasing (1.5 C, 2 C), the consistency is generally becoming higher. Author Contributions: X.Y.: Formal analysis, visualization, writing—original draft. T.W.: formal analysis, visualization, writing—original draft. M.D.: formal analysis, writing—original draft. Y.W.: formal analysis, methodology. W.S.: formal analysis, methodology. Q.Z.: methodology, software, vali- dation. B.H.: funding acquisition, supervision, conceptualization, methodology, validation, writing— review and editing. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Natural Science Foundation of China grant number 42171121, 41701059. And The APC was funded by 42171121. Data Availability Statement: All data and the script of the whole processes are available through an email request to the authors. Acknowledgments: This work was funded by the Natural Science Foundation of China (42171121, 41701059). The authors gratefully acknowledge data availability from the Programme for Monitoring of the Greenland Ice Sheet (PROMICE) and the Greenland Analogue Project (GAP) provided by the Geological Survey of Denmark and Greenland (GEUS) at http://www.promice.dk (accessed on 6 October 2021). Conflicts of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References 1. Steffen, K.; Box, J. Surface climatology of the Greenland ice sheet: Greenland climate network 1995–1999. J. Geophys. Res.-Atmos. 2001, 106, 33951–33964. [CrossRef] 2. Mouginot, J.; Rignot, E.; Bjork, A.A.; van den Broeke, M.; Millan, R.; Morlighem, M.; Noel, B.; Scheuchl, B.; Wood, M. Forty-six years of Greenland Ice Sheet mass balance from 1972 to 2018. Proc. Natl. Acad. Sci. USA 2019, 116, 9239–9244. [CrossRef] [PubMed] 3. Blunden, J.; Arndt, D. State of the Climate in 2019. Bull. Am. Meteorol. Soc. 2020, 101, S1–S429. [CrossRef] 4. Hanna, E.; Huybrechts, P.; Steffen, K.; Cappelen, J.; Huff, R.; Shuman, C.; Irvine-Fynn, T.; Wise, S.; Griffiths, M. Increased runoff from melt from the Greenland Ice Sheet: A response to global warming. J. Clim. 2008, 21, 331–341. [CrossRef] 5. Tedesco, M.; Serreze, M.; Fettweis, X. Diagnosing the extreme surface melt event over southwestern Greenland in 2007. Cryosphere 2008, 2, 159–166. [CrossRef] 6. Mernild, S.H.; Mote, T.L.; Liston, G.E. Greenland ice sheet surface melt extent and trends: 1960–2010. J. Glaciol. 2011, 57, 621–628. [CrossRef] 7. Nghiem, S.V.; Hall, D.K.; Mote, T.L.; Tedesco, M.; Albert, M.R.; Keegan, K.; Shuman, C.A.; DiGirolamo, N.E.; Neumann, G. The extreme melt across the Greenland ice sheet in 2012. Geophys. Res. Lett. 2012, 39, L20502. [CrossRef] 8. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Box, J.E.; Koenig, L.S. Variability in the surface temperature and melt extent of the Greenland ice sheet from MODIS. Geophys. Res. Lett. 2013, 40, 2114–2120. [CrossRef] 9. Hanna, E.; Fettweis, X.; Mernild, S.H.; Cappelen, J.; Ribergaard, M.H.; Shuman, C.A.; Steffen, K.; Wood, L.; Mote, T.L. Atmospheric and oceanic climate forcing of the exceptional Greenland ice sheet surface melt in summer 2012. Int. J. Climatol. 2014, 34, 1022–1037. [CrossRef] Land 2022, 11, 593 13 of 14 10. Rignot, E.; Velicogna, I.; van den Broeke, M.R.; Monaghan, A.; Lenaerts, J. Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophys. Res. Lett. 2011, 38, L05503. [CrossRef] 11. Enderlin, E.M.; Howat, I.M. Submarine melt rate estimates for floating termini of Greenland outlet glaciers (2000–2010). J. Glaciol. 2013, 59, 67–75. [CrossRef] 12. Fettweis, X.; Hanna, E.; Lang, C.; Belleflamme, A.; Erpicum, M.; Gallee, H. Important role of the mid-tropospheric atmospheric circulation in the recent surface melt increase over the Greenland ice sheet. Cryosphere 2013, 7, 241–248. [CrossRef] 13. Wouters, B.; Bamber, J.L.; van den Broeke, M.R.; Lenaerts, J.T.M.; Sasgen, I. Limits in detecting acceleration of ice sheet mass loss due to climate variability. Nat. Geosci. 2013, 6, 613–616. [CrossRef] 14. Comiso, J.C.; Yang, J.; Honjo, S.; Krishfield, R.A. Detection of change in the Arctic using satellite and in situ data. J. Geophys. Res. Ocean. 2003, 108, C12. [CrossRef] 15. Wang, X.J.; Key, J.R. Recent trends in arctic surface, cloud, and radiation properties from space. Science 2003, 299, 1725–1728. [CrossRef] 16. Wang, X.J.; Key, J.R. Arctic surface, cloud, and radiation properties based on the AVHRR Polar Pathfinder dataset. Part I: Spatial and temporal characteristics. J. Clim. 2005, 18, 2558–2574. [CrossRef] 17. Wang, X.J.; Key, J.R. Arctic surface, cloud, and radiation properties based on the AVHRR Polar Pathfinder dataset. Part II: Recent trends. J. Clim. 2005, 18, 2575–2593. [CrossRef] 18. Mote, T.L. Greenland surface melt trends 1973–2007: Evidence of a large increase in 2007. Geophys. Res. Lett. 2007, 34, L22507. [CrossRef] 19. van den Broeke, M.; Bamber, J.; Ettema, J.; Rignot, E.; Schrama, E.; van de Berg, W.J.; van Meijgaard, E.; Velicogna, I.; Wouters, B. Partitioning Recent Greenland Mass Loss. Science 2009, 326, 984–986. [CrossRef] 20. Reddy, S. Comments on IPCC’s 24th September 2019 Report on “The Ocean and Cryosphere in a Changing Climate: Summary for Policy Makers”. Acta Sci. Agric 2019, 3, 16–19. [CrossRef] 21. Pörtner, H.-O.; Roberts, D.C.; Masson-Delmotte, V.; Zhai, P.; Tignor, M.; Poloczanska, E.; Weyer, N. The ocean and cryosphere in a changing climate. IPCC Spec. Rep. Ocean Cryosphere A Chang. Clim. 2019. Available online: https://www.ipcc.ch/site/assets/ uploads/sites/3/2019/12/02_SROCC_FM_FINAL.pdf (accessed on 15 March 2022). 22. Muntjewerf, L.; Petrini, M.; Vizcaino, M.; da Silva, C.E.; Sellevold, R.; Scherrenberg, M.D.W.; Thayer-Calder, K.; Bradley, S.L.; Lenaerts, J.T.M.; Lipscomb, W.H.; et al. Greenland Ice Sheet Contribution to 21st Century Sea Level Rise as Simulated by the Coupled CESM2.1-CISM2.1. Geophys. Res. Lett. 2020, 47, e2019GL086836. [CrossRef] 23. Calov, R.; Beyer, S.; Greve, R.; Beckmann, J.; Willeit, M.; Kleiner, T.; Rueckamp, M.; Humbert, A.; Ganopolski, A. Simulation of the future sea level contribution of Greenland with a new glacial system model. Cryosphere 2018, 12, 3097–3121. [CrossRef] 24. Rueckamp, M.; Greve, R.; Humbert, A. Comparative simulations of the evolution of the Greenland ice sheet under simplified Paris Agreement scenarios with the models SICOPOLIS and ISSM. Polar Sci. 2019, 21, 14–25. [CrossRef] 25. Bell, R.E. The role of subglacial water in ice-sheet mass balance. Nat. Geosci. 2008, 1, 297–304. [CrossRef] 26. van den Broeke, M.R.; Enderlin, E.M.; Howat, I.M.; Munneke, P.K.; Noel, B.P.Y.; van de Berg, W.J.; van Meijgaard, E.; Wouters, B. On the recent contribution of the Greenland ice sheet to sea level change. Cryosphere 2016, 10, 1933–1946. [CrossRef] 27. Smith, L.C.; Chu, V.W.; Yang, K.; Gleason, C.J.; Pitcher, L.H.; Rennermalm, A.K.; Legleiter, C.J.; Behar, A.E.; Overstreet, B.T.; Moustafa, S.E.; et al. Efficient meltwater drainage through supraglacial streams and rivers on the southwest Greenland ice sheet. Proc. Natl. Acad. Sci. USA 2015, 112, 1001–1006. [CrossRef] 28. Hall, D.K.; Cullather, R.I.; DiGirolamo, N.E.; Comiso, J.C.; Medley, B.C.; Nowicki, S.M. A Multilayer Surface Temperature, Surface Albedo, and Water Vapor Product of Greenland from MODIS. Remote Sens. 2018, 10, 555. [CrossRef] 29. Miller, N.B.; Shupe, M.D.; Cox, C.J.; Noone, D.; Persson, P.O.G.; Steffen, K. Surface energy budget responses to radiative forcing at Summit, Greenland. Cryosphere 2017, 11, 497–516. [CrossRef] 30. Noel, B.; van de Berg, W.J.; van Wessem, J.M.; van Meijgaard, E.; van As, D.; Lenaerts, J.T.M.; Lhermitte, S.; Munneke, P.K.; Smeets, C.J.P.P.; van Ulft, L.H.; et al. Modelling the climate and surface mass balance of polar ice sheets using RACMO2-Part 1: Greenland (1958–2016). Cryosphere 2018, 12, 811–831. [CrossRef] 31. Noel, B.; van de Berg, W.J.; Lhermitte, S.; van den Broeke, M.R. Rapid ablation zone expansion amplifies north Greenland mass loss. Sci. Adv. 2019, 5, eaaw0123. [CrossRef] [PubMed] 32. Huai, B.; van den Broeke, M.R.; Reijmer, C.H. Long-term surface energy balance of the western Greenland Ice Sheet and the role of large-scale circulation variability. Cryosphere 2020, 14, 4181–4199. [CrossRef] 33. Nielsen-Englyst, P.; Hoyer, J.L.; Madsen, K.S.; Tonboe, R.T.; Dybkjaer, G.; Skarpalezos, S. Deriving Arctic 2 m air temperatures over snow and ice from satellite surface temperature measurements. Cryosphere 2021, 15, 3035–3057. [CrossRef] 34. Shuman, C.A.; Steffen, K.; Box, J.E.; Stearns, C.R. A dozen years of temperature observations at the summit: Central Greenland automatic weather stations 1987–1999. J. Appl. Meteorol. 2001, 40, 741–752. [CrossRef] 35. Ahlstrom, A.P.; Gravesen, P.; Andersen, S.B.; van As, D.; Citterio, M.; Fausto, R.S.; Nielsen, S.; Jepsen, H.F.; Kristensen, S.S.; Christensen, E.L.; et al. A new programme for monitoring the mass loss of the Greenland ice sheet. Geol. Surv. Den. Greenl. Bull. 2008, 15, 61–64. [CrossRef] 36. van As, D.; Fausto, R.S.; Team, P.P. Programme for Monitoring of the Greenland Ice Sheet (PROMICE): First temperature and ablation records. Geol. Surv. Den. Greenl. Bull. 2011, 23, 73–76. [CrossRef] Land 2022, 11, 593 14 of 14 37. Fausto, R.S.; Van As, D.; Ahlstrom, A.P.; Citterio, M. Assessing the accuracy of Greenland ice sheet ice ablation measurements by pressure transducer. J. Glaciol. 2012, 58, 1144–1150. [CrossRef] 38. Box, J.E.; Rinke, A. Evaluation of Greenland ice sheet surface climate in the HIRHAM regional climate model using automatic weather station data. J. Clim. 2003, 16, 1302–1319. [CrossRef] 39. Leeson, A.A.; Van Wessem, J.M.; Ligtenberg, S.R.M.; Shepherd, A.; Van Den Broeke, M.R.; Killick, R.; Skvarca, P.; Marinsek, S.; Colwell, S. Regional climate of the Larsen B embayment 1980–2014. J. Glaciol. 2017, 63, 683–690. [CrossRef] 40. Medley, B.; Joughin, I.; Das, S.B.; Steig, E.J.; Conway, H.; Gogineni, S.; Criscitiello, A.S.; McConnell, J.R.; Smith, B.E.; van den Broeke, M.R.; et al. Airborne-radar and ice-core observations of annual snow accumulation over Thwaites Glacier, West Antarctica confirm the spatiotemporal variability of global and regional atmospheric models. Geophys. Res. Lett. 2013, 40, 3649–3654. [CrossRef] 41. Hall, D.K.; Box, J.E.; Casey, K.A.; Hook, S.J.; Shuman, C.A.; Steffen, K. Comparison of satellite-derived and in-situ observations of ice and snow surface temperatures over Greenland. Remote Sens. Environ. 2008, 112, 3739–3749. [CrossRef] 42. Fausto, R.S.; van As, D.; Mankoff, K.D.; Vandecrux, B.; Citterio, M.; Ahlstrom, A.P.; Andersen, S.B.; Colgan, W.; Karlsson, N.B.; Kjeldsen, K.K.; et al. Programme for Monitoring of the Greenland Ice Sheet (PROMICE) automatic weather station data. Earth Syst. Sci. Data 2021, 13, 3819–3845. [CrossRef] 43. Ryan, J.; Smith, L.; Van As, D.; Cooley, S.; Cooper, M.; Pitcher, L.; Hubbard, A. Greenland Ice Sheet surface melt amplified by snowline migration and bare ice exposure. Sci. Adv. 2019, 5, eaav3738. [CrossRef] [PubMed] 44. Barnes, W.L.; Pagano, T.S.; Salomonson, V.V. Prelaunch characteristics of the Moderate Resolution Imaging Spectroradiometer (MODIS) on EOS-AM1. IEEE Trans. Geosci. Remote Sens. 1998, 36, 1088–1100. [CrossRef] 45. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Key, J.R.; Koenig, L.S. A Satellite-Derived Climate-Quality Data Record of the Clear-Sky Surface Temperature of the Greenland Ice Sheet. J. Clim. 2012, 25, 4785–4798. [CrossRef] 46. Key, J.; Haefliger, M. Arctic ice surface temperature retrieval from AVHRR thermal channels. J. Geophys. Res. Atmos. 1992, 97, 5885–5893. [CrossRef] 47. Key, J.R.; Collins, J.B.; Fowler, C.; Stone, R.S. High-latitude surface temperature estimates from thermal satellite data. Remote Sens. Environ. 1997, 61, 302–309. [CrossRef] 48. Hall, D.K.; Williams, R.S., Jr.; Luthcke, S.B.; Digirolamo, N.E. Greenland ice sheet surface temperature, melt and mass loss: 2000–2006. J. Glaciol. 2008, 54, 81–93. [CrossRef] 49. Stroeve, J.; Steffen, K. Variability of AVHRR-derived clear-sky surface temperature over the Greenland ice sheet. J. Appl. Meteorol. 1998, 37, 23–31. [CrossRef] 50. Tedesco, M.; Fettweis, X.; van den Broeke, M.R.; van de Wal, R.S.W.; Smeets, C.J.P.P.; van de Berg, W.J.; Serreze, M.C.; Box, J.E. The role of albedo and accumulation in the 2010 melting record in Greenland. Environ. Res. Lett. 2011, 6, 014005. [CrossRef] 51. Gałas, ´ A. Impact of volcanic eruptions on the environment and climatic conditions in the area of Poland (Central Europe). Earth-Sci. Rev. 2016, 162, 58–64. [CrossRef] 52. Abdalati, W.; Steffen, K. Greenland ice sheet melt extent: 1979–1999. J. Geophys. Res. -Atmos. 2001, 106, 33983–33988. [CrossRef] 53. Neff, W.; Compo, G.P.; Ralph, F.M.; Shupe, M.D. Continental heat anomalies and the extreme melting of the Greenland ice surface in 2012 and 1889. J. Geophys. Res. -Atmos. 2014, 119, 6520–6536. [CrossRef] 54. Bennartz, R.; Shupe, M.D.; Turner, D.D.; Walden, V.P.; Steffen, K.; Cox, C.J.; Kulie, M.S.; Miller, N.B.; Pettersen, C. July 2012 Greenland melt extent enhanced by low-level liquid clouds. Nature 2013, 496, 83–86. [CrossRef] [PubMed] 55. Cullather, R.I.; Andrews, L.C.; Croteau, M.J.; Digirolamo, N.E.; Hall, D.K.; Lim, Y.-K.; Loomis, B.D.; Shuman, C.A.; Nowicki, S.M.J. Anomalous Circulation in July 2019 Resulting in Mass Loss on the Greenland Ice Sheet. Geophys. Res. Lett. 2020, 47, e2020GL087263. [CrossRef] 56. Tedesco, M.; Fettweis, X. Unprecedented atmospheric conditions (1948–2019) drive the 2019 exceptional melting season over the Greenland ice sheet. Cryosphere 2020, 14, 1209–1223. [CrossRef] 57. Hearty, T.J., III; Lee, J.N.; Wu, D.L.; Cullather, R.; Blaisdell, J.M.; Susskind, J.; Nowicki, S.M.J. Intercomparison of Surface Temperatures from AIRS, MERRA, and MERRA-2 with NOAA and GC-Net Weather Stations at Summit, Greenland. J. Appl. Meteorol. Climatol. 2018, 57, 1231–1245. [CrossRef] 58. Cullather, R.I.; Nowicki, S.M.J. Greenland Ice Sheet Surface Melt and Its Relation to Daily Atmospheric Conditions. J. Clim. 2018, 31, 1897–1919. [CrossRef] 59. Sasgen, I.; Wouters, B.; Gardner, A.S.; King, M.D.; Tedesco, M.; Landerer, F.W.; Dahle, C.; Save, H.; Fettweis, X. Return to rapid ice loss in Greenland and record loss in 2019 detected by the GRACE-FO satellites. Commun. Earth Environ. 2020, 1, 8. [CrossRef] 60. Kimball, J.S.; Du, J.; Meierbachtol, T.W.; Kim, Y.; Johnson, J.V. Comparing Greenland Ice Sheet Melt Variability from Different Satellite Passive Microwave Remote Sensing Products Over a Common 5-year Record. Front. Earth Sci. 2021, 9, 654220. [CrossRef]

Journal

Land – Multidisciplinary Digital Publishing Institute

Published: Apr 19, 2022

Keywords: MODIS; ice surface temperature (IST); validation; Greenland ice sheet (GrIS); variability

There are no references for this article.