Scaling Surface Fluxes from Tower Footprint to Global Model Pixel Scale Using Multi-Satellite Data Fusion

Friday, 19 December 2014: 10:35 AM
Martha C. Anderson1, Christopher Hain2, Feng Gao3, Kathryn A Semmens3, Yun Yang3, Mitchell A Schull1, Theresa Ring4, William P Kustas3 and Joseph G Alfieri3, (1)USDA ARS, Pendleton, OR, United States, (2)Earth System Science Interdisciplinary Center, COLLEGE PARK, MD, United States, (3)USDA ARS, Beltsville, MD, United States, (4)Oregon State University, Corvallis, OR, United States
There is a fundamental challenge in evaluating performance of global land-surface and climate modeling systems, given that few in-situ observation sets adequately sample surface conditions representative at the global model pixel scale (10-100km). For example, a typical micrometeorological flux tower samples a relatively small footprint ranging from 100m to 1km, depending on tower height and environmental conditions. There is a clear need for diagnostic tools that can effectively bridge this gap in scale, and serve as a means of benchmarking global prognostic modeling systems under current conditions. This paper discusses a multi-scale energy balance modeling system (the Atmosphere-Land Exchange Inverse model and disaggregation utility: ALEXI/DisALEXI) that fuses flux maps generated with thermal infrared (TIR) imagery collected by multiple satellite platforms to estimate daily surface fluxes from field to global scales. These diagnostic assessments, with land-surface temperature (LST) as the primary indicator of surface moisture status, operate under fundamentally different constraints than prognostic land-surface models based on precipitation and water balance, and therefore can serve as a semi-independent benchmark. Furthermore, LST can be retrieved from TIR imagery over a broad range of spatiotemporal resolutions: from several meters (airborne systems; periodically) to ~100m (Landsat; bi-weekly) to 1km (Moderate Resolution Imaging Spectroradiometer - MODIS; daily) to 3-10km (geostationary; hourly). Applications of ALEXI/DisALEXI to flux sites within the US and internationally are described, evaluating daily evapotranspiration retrievals generated at 30m resolution. Annual timeseries of maps at this scale can be useful for better understanding local heterogeneity in the tower vicinity and dependences of observed fluxes on wind direction. If reasonable multi-year performance is demonstrated at the tower footprint scale for flux networks such as the National Ecological Observatory Network (NEON), the aggregated output from such diagnostic modeling systems can provide valuable scale-appropriate benchmarks for coarser-resolution models at the tower sites, and some degree of independent verification in areas with sparse or inexistent ground-based measurements.