Landcare Research has performed the first national-scale assessment of the uncertainty associated with estimates of carbon in New Zealand forests.
This work has given Ministry for the Environment (MfE) confidence in the accuracy of our national carbon estimates and will ensure New Zealand meets the United Nations Framework Convention on Climate Change (UNFCCC) reporting guidelines.
The aggregation of carbon stock data is a good example of ecologists’ increasing use of large datasets with highly derived metrics to examine trends at national, regional or even global scales.
Scientists have aggregated data on the change in carbon (based on data obtained from a network of plots) across countries or even biomes (e.g. the tropics) over several years and then used it to make claims about whether forests are net sources or sinks of carbon.
The challenge is that any error in plot measurements or model uncertainty can lead to false conclusions. It is essential to show the correct level of uncertainty in derived metrics so that management implications and policy decisions can be assessed with the appropriate level of confidence.
We assessed the uncertainty associated with measurements of carbon in plots (any error in measuring tree heights, diameters, etc.) and uncertainty with the models used to scale these individual tree measurements into plot and then national-scale carbon estimates.
We found that above-ground carbon stock estimates had 1% greater uncertainty with error propagation than without. Carbon change estimates were much more sensitive to measurement and model uncertainty, with a 43% increase in uncertainty.
Our findings show that national-scale plot-based estimates of carbon stock and carbon change are relatively robust to measurement error and model uncertainty but that carbon stock change between two measurement periods needs to be larger than we previously thought in order for the current national plot network to detect it.
MfE has commissioned us to run our uncertainty estimation alongside the national calculations of Kyoto Protocol commitment period 1 carbon stock change for natural forests. This meets the recommendations of the UNFCCC Good Practice Guidance and will be the first time this type of uncertainty has been provided in New Zealand reports to the UNFCCC.
MfE can have increased confidence that the carbon estimates from the existing Land Use Change and Carbon Analysis System (LUCAS) plot network have relatively little bias and they know the size of stock change they can confidently detect.