TellUs
TellUs is a comprehensive Benchmarking framework developed to evaluate dynamic global vegetation models (DGVMs) and land surface models (LSMs)
The purpose of TellUs is to provide flexible and comprehensive benchmarking, evaluating simulated output against data and or other simulation output. This multi-model framework offers two active versions, TellMeGlobal and TellMeEurope.
TellMeGlobal focus on global simulations and offers evaluation of Spatial patterns, Temporal trends and Seasonal analysis. Benchmarks include PNV Biomes, GOSIF GPP, MODIS LAI, Global burnt area, Pan Forest biomass, Regrowth, FLUXNET and Global Biomass. TellMeEurope is tailored for European scale data input and simulation output. The report handles both spatial (gridded) and temporal (time-series) evaluation. Spatial benchmarks include NFI derived woody biomass growth, stem density, QDBH, and satellite derived GPP. Additionally include plant functional type comparisons. Temporal benchmarks include LAI, biomass, stem density, QDBH and DBH from the PROFOUND forest management sites network. Also benchmarking eddy covariance flux sites from the ICOS & FLUXNET network (covering gross primary productivity, net ecosystem exchange, total ecosystem respiration, evapotranspiration). As of this version the available file formats compatible with the evaluation structure are LPJ-GUESS (.out), NetCDF (.nc), aDGVM (.nc), aDGVM2 (.nc), ICOS, FLUXNET and SITE.
Some examples:
Spatial Benchmark(s) displaying the difference in annual GPP between a LPJ-GUESS simulation and a satellite-derived data product. The range in colour represents an underestimation of the simulated output while blue represents an overestimated GPP. Additionally, the Max, Mean, Min difference of the map as well as the Data mean is displayed in the subtitle. The density plot display how close the combined mapping between your input is to the 1:1 line and how we under/overestimate.
Temporal Benchmark(s) displaying the daily GPP over a couple of stations. Clearly showing how closely the simulated GPP and ICOS flux station data follow each other over time. The scatterplot displays the seasonal mean GPP in relation to the 1:1 line for simulated and measured daily GPP. Additionally, each seasonal point display variability within that season.
Project Contributors: Karl Piltz, Matthew Forest (BiK-F, Germany), Stefan Olin, Thomas Pugh
Benchmarking functionality for DGVMs/LSMs built on DGVMTools.