Satellite Microwave Remote Sensing of Boreal-Arctic Land Surface State and Meteorology from AMSR-E

High latitude regions are undergoing significant climate-related change and represent an integral component of the Earth’s climate system. Near-surface vapor pressure deficit, soil temperature, and soil moisture are essential state variables for monitoring high latitude climate and estimating the re...

Full description

Bibliographic Details
Main Author: Jones, Lucas Alan
Format: Thesis
Language:unknown
Published: University of Montana 2007
Subjects:
Online Access:https://scholarworks.umt.edu/etd/861
https://scholarworks.umt.edu/context/etd/article/1880/viewcontent/msthesis_laj_final.pdf
Description
Summary:High latitude regions are undergoing significant climate-related change and represent an integral component of the Earth’s climate system. Near-surface vapor pressure deficit, soil temperature, and soil moisture are essential state variables for monitoring high latitude climate and estimating the response of terrestrial ecosystems to climate change. Methods are developed and evaluated to retrieve surface soil temperature, daily maximum/minimum air temperature, and land surface wetness information from the EOS Advanced Microwave Scanning Radiometer (AMSR-E) on the Aqua satellite for eight Boreal forest and Arctic tundra biophysical monitoring sites across Alaska and northern Canada. Daily vapor pressure deficit is determined by employing AMSR-E daily maximum/minimum air temperature retrievals. The seasonal pattern of microwave emission and relative accuracy of the estimated land surface state are influenced strongly by landscape properties including the presence of open water, vegetation type and seasonal phenology, snow cover and freeze-thaw transitions. Daily maximum/minimum air temperature is retrieved with RMSEs of 2.88 K and 2.31 K, respectively. Soil temperature is retrieved with RMSE of 3.1 K. Vapor pressure deficit (VPD) is retrieved to within 427.9 Pa using thermal information from AMSR-E. AMSR-E thermal information imparted 27% of the overall error in VPD estimation with the remaining error attributable to underlying algorithm assumptions. Land surface wetness information derived from AMSR-E corresponded with soil moisture observations and simple soil moisture models at locations with tundra, grassland, and mixed -forest/cropland land covers (r = 0.49 to r = 0.76). AMSR-E 6.9 GHz land surface wetness showed little correspondence to soil moisture observation or model estimates at locations with > 20% open water and > 5 m2 m-2 Leaf Area Index, despite efforts to remove the impact of open water and vegetation biomass. Additional information on open water fraction and vegetation phenology derived from ...