Quantifying the prediction accuracy of a 1-D SVAT model at a range of ecosystems in the USA and Australia: evidence towards its use as a tool to study Earth's system interactions

摘要

This paper describes the validation of the SimSphere SVAT (Soil–Vegetation–Atmosphere Transfer) model conducted ata range of US andAustralian ecosystem types. Specific focus was given to examining the models' abilityin predicting shortwave incoming solar radiation (R_g), net radiation(R_net), latent heat (LE), sensible heat (H), air temperature at 1.3 m(T_{air 1.3 m}) and air temperature at 50 m (T_{air 50 m}). Modelpredictions were compared against corresponding in situ measurements acquired for atotal of 72 selected days of the year 2011 obtained from eight sites belongingto the AmeriFlux (USA) and OzFlux (Australia) monitoring networks. Selectedsites were representative of a variety of environmental, biome and climaticconditions, to allow for the inclusion of contrasting conditions in themodel evaluation.Overall, results showed a good agreement between the model predictions andthe in situ measurements, particularly so for the R_g, R_net, T_air1.3 m and T_air 50 m parameters. The simulated R_g parameter exhibiteda root mean square deviation (RMSD) within 25 % of the observed fluxes for58 of the 72 selected days, whereas an RMSD within~ 24 % of the observed fluxes was reported for the R_netparameter for all days of study (RMSD = 58.69 W m?2). A systematicunderestimation of R_g and R_net (mean bias error (MBE) = ?19.48and ?16.46 W m?2) was also found. Simulations for theT_air 1.3 m and T_air 50 m showed good agreement with thein situ observations, exhibiting RMSDs of 3.23 and3.77 °C (within ~ 15 and ~ 18 %of the observed) for all days of analysis, respectively. Comparable, yetslightly less satisfactory simulation accuracies were exhibited for the Hand LE parameters (RMSDs = 38.47 and 55.06 W m?2, ~ 34 and ~ 28 % of theobserved). Highest simulation accuracies were obtained for the open woodlandsavannah and mulga woodland sites for most of the compared parameters.The Nash–Sutcliffe efficiency index for all parameters ranges from 0.720 to0.998, suggesting a very good model representation of the observations.To our knowledge, this study presents the most detailed evaluation ofSimSphere done so far, and the first validation of it conductedin Australian ecosystem types. Findings are important and timely, given theexpanding use of the model both as an educational and research tool today.This includes ongoing research by different space agencies examining itssynergistic use with Earth observation data towards the development ofglobal operational products.