Summary Report No. 30


Comparing global models of terrestrial net primary productivity (NPP): Overview and key results

W. Cramer, D. W. Kicklighter, A. Bondeau, B. Moore III, G. Churkina, A. Ruimy, A. Schloss, participants of "Potsdam ’95" (October 1997)

Seventeen models of terrestrial biogeochemistry have been compared with respect to total, regional, annual and seasonal fluxes of net primary productivity (NPP). The comparison, sponsored by IGBP-GAIM/DIS/GCTE, used standardised input variables wherever possible and was carried out through two international workshops, as well as subsequent distributed work in the participating groups. The models differed widely in complexity and purpose, but could be grouped in three major categories: Satellite-based models that use data from the NOAA/AVHRR sensor as their major input stream (CASA, GLO-PEM, SDBM, SIB2 and TURC), models for seasonal fluxes that use a prescribed vegetation structure (BIOME-BGC, CARAIB 2.1, CENTURY 4.0, FBM 2.2, HRBM 3.0, KGBM, PLAI 0.2, SILVAN 2.2 and TEM 4.0), and models that simulate structure and fluxes (BIOME3, DOLY and HYBRID 3.0). Results from the comparison are summarised in this overview and in six companion papers.

The model simulations resulted in a range of total NPP values (39.9 - 80.5 Pg C yr-1), however this range decreased considerably when two obvious outliers (which produced extreme results due to calibration problems connected with the comparison, had been removed. The broad global pattern of NPP and the relationship of annual NPP to the major climatic variables coincided in many areas. Systematic biases due to the fundamental modeling strategy could not be found, but the inclusion of nutrient constraints seemed to produce lower values. The regional and global sensitivity of fluxes against the simulation method for the water balance was clearly illustrated. Seasonal variation among models was higher, both globally and locally, and this provides several indications for specific deficiencies in the models. A decomposition of annual NPP for most models that do not use remote sensing data into absorbed photosynthetically active radiation (APAR) and light use efficiency (LUE), with subsequent comparison of these values against those used by the remote sensing models, provided a negative relationship between total APAR and LUE across these models. This may indicate that the models are calibrated to achieve 'commonly accepted values' of the total flux, while doing so with widely differing spatial and seasonal patterns.