Leaf Area Index in Earth System Models: Evaluation and Projections

Abstract:

The latest generation of earth system models simulate leaf area index, as well as provide projections of leaf area index (LAI) in the future. Here we use satellite measurements of LAI to look at the ability of the current generation of models to simulate the seasonal cycle and interannual variability of LAI. In addition, we look at observed temperature and precipitation relationships with interannual variability in LAI, and consider whether the models are able to correctly simulate these relationships. We use this analysis to also examine how useful particular metrics are for inter-model comparisons and model skill evaluation. We propose that good metrics for evaluating coupled earth system models are those that are relatively insensitive to the meteorology used to force them. In addition, it is valuable to have skill metrics that correlate with future projections, which would allow us to eliminate some future projections as less likely. Depending on the region of interest (tropics, mid-latitudes or high latitudes), we show that different metrics are of the most use. For future projections over the whole globe and in the tropics the most important metrics are related to current LAI correlations with time, as well as the magnitude of the mean LAI. However, several metrics are shown to have significant correlations with future projections, inhibiting our ability to use the correlation of skill metrics with future projections to improve future projections.