MTSAT-1R land surface temperature retrievals and their potential role as observational constraints on surface energy balance estimation for Australia

[摘要] Geostationary satellite observations have the potential to provide new insight into the sub-daily dynamics of energy and water cycles. To be of direct use, thermal brightness temperature observations require processing to land surface temperature (LST) values. The most efficient way of performing this over the large spatial coverage and data volumes of digital imagery is via the split-window algorithm (SWA).This work applied a number of published SWA’s to the thermal brightness temperature data from the Japanese Advanced Meteorological Imagery (JAMI) onboard the MTSAT-1R geostationary satellite, with the intention of selecting the best SWA for routine application. Performance was assessed using the state-of-the-art daily MODIS LST product for coincident image pairs. Results showed that there is no standout SWA and, moreover, SWA’s representing the extremes of sophistication in terms of number of algorithm inputs were shown to yield similar results.Analysis of the SWA’s performance revealed a distinct day/night difference. Night time JAMI retrievals had on average ~1 K bias and overall RMS difference2 K. Daytime retrievals were considerably poorer with individual bias for some SWA’s5 K. Generally, it was concluded that the best expected error for JAMI-derived LST was on average ~3 K.A cursory exploration into the use of JAMI-derived LST in radiation balance estimation and as a surface energy balance constraint was presented. Modelled estimates of net radiation were obtained using JAMI LST data in the upwelling long-wave radiation term. Comparison with on-site measured net radiation revealed an overestimate of the modelled value by ~20 W m-2 and an overall agreement of ~50 W m-2. JAMI LST data were used as an observational constraint on a simple two-layer surface energy balance (SEB) model. By sequentially adjusting the stomatal resistance parameter in the SEB model, observed and modelled estimates of LST were brought into agreement at the time of image acquisition. Resulting model estimates of latent and sensible heat flux mimicked what would one expect to see in reality. More work is needed to verify the usefulness of JAMI LST for this role.Recommendations for further developments and research tasks include:* Extending the analysis of JAMI performance in LST and surface energy balance estimation to cover a longer time periods, with focused comparisons at strategically selected sites (e.g. flux towers).* Examining coincident bright temperature image pairs of AVHRR and/or MODIS and JAMI data to determine (a) the nature of the spectral response between the two sensors, (b) the radiometric quality of the data, and (c) cross-calibration relationships. * Supporting investigations into the development of a JAMI cloud and cloud shadow masking algorithm. * Exploring the use of JAMI-derived LST and other geospatial data set (e.g. soil moisture, microwave radiances, or vegetation indices) in a multiple-constraints model-data fusion framework for coupled water-energy balance model estimation.Ultimately the constraint that derived LST observations will impart on surface energy balance will lead to improved soil water balance modelling (e.g. profile soil moisture estimation) with benefits for runoff estimation.

[发布日期] 2009-06-20 [发布机构] CSIRO

[学科分类] 地球科学(综合)