Final published version
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - A comparison of supraglacial lake observations derived from MODIS imagery at the western margin of the Greenland ice sheet
AU - Leeson, Amber A.
AU - Shepherd, Andrew
AU - Sundal, Aud V.
AU - Johansson, A. Malin
AU - Selmes, Nick
AU - Briggs, Kate
AU - Hogg, Anna E.
AU - Fettweis, Xavier
PY - 2013/12/1
Y1 - 2013/12/1
N2 - Supraglacial lakes (SGLs) affect the dynamics of the Greenland ice sheet by storing runoff and draining episodically. We investigate the evolution of SGLs as reported in three datasets, each based on automated classification of satellite imagery. Although the datasets span the period 2001-10, there are differences in temporal sampling, and only the years 2005-07 are common. By subsampling the most populous dataset, we recommend a sampling frequency of one image per 6.5 days in order to minimize uncertainty associated with poor temporal sampling. When compared with manual classification of satellite imagery, all three datasets are found to omit a sizeable (29, 48 and 41%) fraction of lakes and are estimated to document the average size of SGLs to within 0.78, 0.48 and 0.95km2. We combine the datasets using a hierarchical scheme, producing a single, optimized, dataset. This combined record reports up to 67% more lakes than a single dataset. During 2005-07, the rate of SGL growth tends to follow the rate at which runoff increases in each year. In 2007, lakes drain earlier than in 2005 and 2006 and remain absent despite continued runoff. This suggests that lakes continue to act as open surface-bed conduits following drainage.
AB - Supraglacial lakes (SGLs) affect the dynamics of the Greenland ice sheet by storing runoff and draining episodically. We investigate the evolution of SGLs as reported in three datasets, each based on automated classification of satellite imagery. Although the datasets span the period 2001-10, there are differences in temporal sampling, and only the years 2005-07 are common. By subsampling the most populous dataset, we recommend a sampling frequency of one image per 6.5 days in order to minimize uncertainty associated with poor temporal sampling. When compared with manual classification of satellite imagery, all three datasets are found to omit a sizeable (29, 48 and 41%) fraction of lakes and are estimated to document the average size of SGLs to within 0.78, 0.48 and 0.95km2. We combine the datasets using a hierarchical scheme, producing a single, optimized, dataset. This combined record reports up to 67% more lakes than a single dataset. During 2005-07, the rate of SGL growth tends to follow the rate at which runoff increases in each year. In 2007, lakes drain earlier than in 2005 and 2006 and remain absent despite continued runoff. This suggests that lakes continue to act as open surface-bed conduits following drainage.
U2 - 10.3189/2013JoG13J064
DO - 10.3189/2013JoG13J064
M3 - Journal article
AN - SCOPUS:84894160211
VL - 59
SP - 1179
EP - 1188
JO - Journal of Glaciology
JF - Journal of Glaciology
SN - 0022-1430
IS - 218
ER -