ALBERT

Description: Surface temperatures on the Greenland Ice Sheet have been studied on the ground, using automatic weather station (AWS) data from the Greenland-Climate Network (GC-Net), and from analysis of satellite sensor data. Using Advanced Very High Frequency Radiometer (AVHRR) weekly surface temperature maps, warming of the surface of the Greenland Ice Sheet has been documented from 1981 to present. We extend and refine this record using higher-resolution Moderate-Resolution Imaging Spectroradiometer (MODIS) data from March 2000 to the present. To permit changes to be observed over time, we are developing a well-characterized monthly climate-data record (CDR) of the "clear-sky" surface temperature of the Greenland Ice Sheet using data from both the Terra and Aqua satellites. We use the MODIS ice-surface temperature (IST) algorithm. Validation of the CDR consists of several facets: 1) comparisons between the Terra and Aqua IST maps; 2) comparisons between ISTs and in-situ measurements; 3) comparisons between ISTs and AWS data; and 4) comparisons of ISTs with surface temperatures derived from other satellite instruments such as the Thermal Emission and Reflection Radiometer. In this work, we focus on 1) and 2) above. Surface temperatures on the Greenland Ice Sheet have been studied on the ground, using automatic weather station (AWS) data from the Greenland-Climate Network (GC-Net), and from analysis of satellite sensor data. Using Advanced Very High Frequency Radiometer (AVHRR) weekly surface temperature maps, warming of the surface of the Greenland Ice Sheet has been documented from 1981 to present. We extend and refine this record using higher-resolution Moderate-Resolution Imaging Spectroradiometer (MODIS) data from March 2000 to the present. To permit changes to be observed over time, we are developing a well-characterized monthly climate-data record (CDR) of the "clear-sky" surface temperature of the Greenland Ice Sheet using data from both the Terra and Aqua satellites. We use the MODIS ice-surface temperature (IST) algorithm. Validation of the CDR consists of several facets: 1) comparisons between the Terra and Aqua IST maps; 2) comparisons between ISTs and in-situ measurements; 3) comparisons between ISTs and AWS data; and 4) comparisons of ISTs with surface temperatures derived from other satellite instruments such as the Thermal Emission and Reflection Radiometer. In this work, we focus on 1) and 2) above. First we provide comparisons between Terra and Aqua swath-based ISTs at approximately 14:00 Local Solar Time, reprojected to 12.5 km polar stereographic cells. Results show good correspondence when Terra and Aqua data were acquired within 2 hrs of each other. For example, for a cell centered over Summit Camp (72.58 N, 38.5 W), the average agreement between Terra and Aqua ISTs is 0.74 K (February 2003), 0.47 K (April 2003), 0.7 K (August 2003) and 0.96 K (October 2003) with the Terra ISTs being generally lower than the Aqua ISTs. More precise comparisons will be calculated using pixel data at the swath level, and correspondence between Terra and Aqua IST is expected to be closer. (Because of cloud cover and other considerations, only a few common cloud-free swaths are typically available for each month for comparison.) Additionally, previous work comparing land-surface temperatures (LSTs) from the standard MODIS LST product and in-situ surface-temperature data at Summit Camp on the Greenland Ice Sheet show that Terra MODIS LSTs are about 3 K lower than in-situ temperatures at Summit Camp, during the winter of 2008-09. This work will be repeated using both Terra and Aqua IST pixel data (in place of LST data). In conclusion, we demonstrate that the uncertainties in the CDR will be well characterized as we work through the various facets of its validation.

Description: We are developing a climate-data record (CDR of daily "clear-sky" ice-surface temperature (IST) of the Greenland Ice Sheet, from 1982 to the present using Advanced Very High Resolution Radiometer (AVHRR) (1982 - present) and Moderate-Resolution Imaging Spectroradiometer (MODIS) data (2000 - present) at a resolution of approximately 5 km. The CDR will be continued in the National Polar-orbiting Operational Environmental Satellite System Visible/Infrared Imager Radiometer Suite era. Two algorithms remain under consideration. One algorithm under consideration is based on the split-window technique used in the Polar Pathfinder dataset (Fowler et al., 2000 & 21007). Another algorithm under consideration, developed by Comiso (2006), uses a single channel of AVHRR data (channel 4) in conjunction with meteorological-station data to account for atmospheric effects and drift between AVHRR instruments. Known issues being addressed in the production of the CDR are: tune-series bias caused by cloud cover (surface temperatures can be different under clouds vs. clear areas) and cross-calibration in the overlap period between AVHRR instruments, and between AVHRR and MODIS instruments. Because of uncertainties, mainly due to clouds (Stroeve & Steffen, 1998; Wang and Key, 2005; Hall et al., 2008 and Koenig and Hall, submitted), time-series of satellite 1S'1" do not necessarily correspond to actual surface temperatures. The CDR will be validated by comparing results with automatic-,",eather station (AWS) data and with satellite-derived surface-temperature products. Regional "clear-sky" surface temperature increases in the Arctic, measured from AVHRR infrared data, range from 0.57+/-0.02 deg C (Wang and Key, 2005) to 0.72+/-0.10 deg C (Comiso, 2006) per decade since the early 1980s. Arctic warming has important implications for ice-sheet mass balance because much of the periphery of the Greenland Ice Sheet is already near 0 deg C during the melt season, and is thus vulnerable to rapid melting if temperatures continue to increase. References

Description: A new multilayer IST-albedo Moderate Resolution Imaging Spectroradiometer (MODIS) product of Greenland was developed to meet the needs of the ice sheet modeling community. The multiple layers of the product enable the relationship between IST and albedo to be evaluated easily. Surface temperature is a fundamental input for dynamical ice sheet models because it is a component of the ice sheet radiation budget and mass balance. Albedo influences absorption of incoming solar radiation. The daily product will combine the existing standard MODIS Collection-6 ice-surface temperature, derived melt maps, snow albedo and water vapor products. The new product is available in a polar stereographic projection in NetCDF format. The product will ultimately extend from March 2000 through the end of 2017.

Description: To quantify the ice-surface temperature (IST) we are developing a climate-data record (CDR) of monthly IST of the Greenland ice sheet, from 1982 to the present using Advanced Very High Resolution Radiometer (AVHRR) and Moderate-Resolution Imaging Spectroradiometer (MODIS) data at 5-km resolution. "Clear-sky" surface temperature increases have been measured from the early 1980s to the early 2000s in the Arctic using AVHRR data, showing increases ranging from 0.57-0.02 (Wang and Key, 2005) to 0.72 0.10 deg C per decade (Comiso, 2006). Arctic warming has implications for ice-sheet mass balance because much of the periphery of the ice sheet is near 0 deg C in the melt season and is thus vulnerable to more extensive melting (Hanna et al., 2008). The algorithm used for this work has a long history of measuring IST in the Arctic with AVHRR (Key and Haefliger, 1992). The data are currently available from 1981 to 2004 in the AVHRR Polar Pathfinder (APP) dataset (Fowler et al., 2000). J. Key1NOAA modified the AVHRR algorithm for use with MODIS (Hall et al., 2004). The MODIS algorithm is now being processed over Greenland. Issues being addressed in the production of the CDR are: time-series bias caused by cloud cover, and cross-calibration between AVHRR and MODIS instruments. Because of uncertainties, time series of satellite ISTs do not necessarily correspond with actual surface temperatures. The CDR will be validated by comparing results with in-situ (see Koenig and Hall, in press) and automatic-weather station data (e.g., Shuman et al., 2001).