Designing systems to monitor carbon stocks in forests and shrublands

Abstract

The United Nations framework convention on climate change (FCCC) commits signatory nations to monitor changes in all fluxes and sinks of carbon, including those related to vegetation and soil. This paper describes a system for monitoring carbon in New Zealand’s forests and shrublands (6.3 and 2.6 million ha, respectively), which was tested on a 60 km-wide transect across the South Island. A 9 km 2 grid was superimposed onto a land-cover map (SPOT imagery) to obtain 62 ground-sampling points. New permanent plots were established at 43 of these points and 18 existing plots were revisited (one site was inaccessible). On each plot, the dimensions of all trees, shrubs and coarse woody debris (CWD) were measured, and these measurements converted to per-hectare C-stocks using regression relationships. C-stocks in fine litter and mineral soil were quantified using data from New Zealand’s national soil-C monitoring system that quantifies national-scale soil-C stocks using a soil-C database stratified by soil type, climate, and land-cover; we used these data to derive values for the transect. The total C-stock in all pools was 290 t/ha in forests and 163 t/ha in shrublands, 58 and 30% of which was living biomass (respectively). The living biomass components had the greatest variances, so precision could be improved by increasing the sample size of trees, particularly large trees. The C-stock of CWD was highly variable among plots, but was a relatively small component of the total C-stock (10% forest and 3% shrubland). For the moment, we know little about the uncertainty associated with measuring C-sequestration (i.e. change in C-stocks over time), but this information will become available when plots are resampled. Our network of permanent plots is designed deliberately to be simple and measurements of interest to conservation biologists, such as species composition and structure are included in the hope of protecting it against inevitable shifts in policy and focus. We discuss methodological issues including generalized allometric relationship, the propagation of errors, and better integration of remote-sensing and ground-based approaches, all of which have been considered in the development of this system.