Combining Remote Sensing and Forest Ecosystem Modeling: An Example Using the Regional HydroEcological Simulation System (RHESSys)

Abstract

Images from airborne or satellite-based remote sensing systems are the only data available for regional and global productivity studies that do not require interpolation or extrapolation. Four categories of image use are identified: image classification, model initialization, model input and model verification. Model initialization using vegetation indices derived from images is discussed using a regional modeling framework, the Regional HydroEcological Simulation System (RHESSys). In this chapter, the sensitivity of RHESSys to soil moisture, and the interrelationships between the soil data theme and the vegetation and climate data themes are illustrated. Improving image transfer functions can increase the quality of vegetation estimates; however, ancillary data (such as topography and soil data) are also needed at appropriate levels of accuracy and precision. An example simulation is provided, which uses vegetation data from two watersheds in western Montana, USA. Results demonstrate the sensitivity of the model to soil data in a wet, dry climate, and indicate the importance of considering the data collection process as an integrated effort guided by modelling requirements and model sensitivity. Additional consideration must be made for validation and collection of independent data for these purposes.