Chapter 15 Terrestrial Net Primary Production (NPP) Estimation Using NOAA Satellite Imagery: Inter-annual Changes between 1982 and 1999

Abstract

The light use efficiency (LUE) approach using the normalized difference vegetation index (NDVI) is the simplest method to estimate terrestrial net primary production (NPP). A simple LUE-based model, which used LUEs of plants, solar radiation, NDVI and stress functions for soil water and air temperature, was applied to estimate global terrestrial NPP between 1982 and 1999. NDVI images were computed using channels 1 and 2 of the Advanced Very High Resolution Radiometer (AVHRR) produced by the Pathfinder AVHRR LAND (PAL) datasets, after a simple radiometric correction. Anomalies in time series of global average monthly NDVI of the original PAL data, which were probably caused by volcanic ashes exposed in the atmosphere by huge eruptions such as Pinatubo, were successfully reduced in the computed NDVI. The solar radiation, air temperature and soil water of NCEP/NCAR Reanalysis data were also used for the NPP estimation. LUEs were determined based on the maximum photosynthetic rate in the literature for 17 land cover types of the IGBP-DIS classification. NPP was also computed using a constant LUE for the global vegetation to understand the effects of multiple LUE-settings. As a result, global NPP was estimated at between 58.5 and 62.6 petagram (Pg) carbon in 1983 and 1998, respectively, for the multiple LUE-settings and between 51.0 and 54.8 Pg carbon in 1983 and 1991, respectively, for the constant LUE-setting. NPP increased about 0.1 Pg carbon annually ( P <0.1) for the case of multiple LUE-settings and about 0.08 Pg carbon annually (not significant) for the case of constant LUE-setting. Carbon absorption by the terrestrial ecosystems in 1980s and 1990s are reported in a literature (Kalnay et al., 1996) based on an analysis of atmospheric CO 2 concentration. The rate of change in carbon absorption observed in that study is similar to that of NPP by our estimation. Inter-annual changes in NPP varied from region to region. Most clear long-term increasing or decreasing trends of NPP appeared in semiarid areas. The inter-annual changes in NDVI seemed to relate most clearly to the inter-annual NPP changes in these areas among parameters used for the NPP estimation, which were NDVI, average monthly air temperature, volumetric soil water content and solar radiation.