Perspectives on combining ecological process models and remotely sensed data

Abstract

A major occupation of the terrestrial remote sensing community is the derivation of spatially comprehensive estimates of biophysical parameters, for example, leaf area index. However, the majority of this work has been attempted without reference to the needs of ecological process studies. While model initialisation is an important task there are other ways in which remote sensing can be combined with ecological models. This paper identifies four alternative strategies: (i) to use remotely sensed data to provide estimates of variables required to drive ecological process models, (ii) to use remotely sensed data to test, validate or verify predictions of ecological process models, (iii) to use remotely sensed data to update or adjust ecological process model predictions and (iv) to use ecological process models to understand remotely sensed data. The objectives of the paper are to review the four strategies by reference to examples. Directions for future work are identified for each strategy, which can be grouped in terms of estimation accuracy, issues of spatial and temporal scale, long term comprehensive datasets and development of new methods and models with the emphasis on increased interaction between scientists from the remote sensing and ecological modelling disciplines.