Land-use change alters New Zealand’s terrestrial carbon budget: uncertainties associated with estimates of soil carbon change between 1990–2000

Abstract

In New Zealand, afforestation and reforestation of grazing land give rise to large (relative to national CO2 emissions) vegetation carbon (C) sinks. These land-use changes may, however, lead to losses of mineral soil C. Full C accounting may, therefore, require including mineral soil C losses if credits are awarded for vegetation C. To monitor soil C stocks and changes, we developed an IPCC-based soil Carbon Monitoring System (CMS) in which New Zealand is stratified by soil type, climate, land-use and an erosivity index (slope ? precipitation). Georeferenced soil C data were used to assign steady-state soil C stocks to various combinations of these factors (cells). We then used a General Linear Model to compare soil C between cells, and derived land-use effects (LUEs) from this analysis that quantify soil C changes that accompany land-use change. These LUEs were used to predict soil C changes resulting from land-use change between 1990?2000. We tested the CMS by comparing predicted soil C stocks, and changes in these stocks, against more detailed soil C data. Overall, soil C estimates obtained from the CMS are consistent with detailed, stratified soil C measurements at specific sites and over larger regions. However, for grazing-land to exotic-forest conversions, estimates of soil C changes are higher and more variable than those based on paired-site studies. Nationally, soil C losses of 0.9 ± 0.4 Tg C yr?1 for all land-use changes over the period 1990?2000 appear likely, with uncertainties arising mainly from estimates of changes in the areas involved, and LUE values for cells with limited soil C data. Changes in soil C in reforested land are likely to be small, but precise area changes, soil C data and detailed paired-site studies are lacking for this key land-use change. By contrast, biomass C accumulation in new exotic plantation forests (afforestation) and native reforestation are 6?9 Tg C yr?1, with C accumulation by afforestation being well quantified. Detailed site studies, process-based modelling and improved area-change estimates are needed to further reduce uncertainties in land-use-change effects on soil C. Our approach could be adapted to countries with country-specific land-use issues different from those in the IPCC default methodology.