Integration of Satellite remote sensing data in ecosystem modelling at local scales: Practices and trends

A review of the ECOPOTENTIAL paper: Integration of Satellite remote sensing data in ecosystem modelling at local scales: Practices and trends– D. Pasetto, S. Arenas-Castro, J. Bustamante, R. Casagrandi, N. Chrysoulakis, A. F. Cord, A. Dittrich, C. Domingo-Marimon, G. El Serafy, A. Karnieli, G. A. Kordelas, I. Manakos, L. Mari, A. Monteiro, E. Palazzi, D. Poursanidis, A. Rinaldo, S. Terzago, A. Ziemba, G. Ziv  Methods in Ecology and Evolution, 9(8), pp 1810-1821


Damiano Pasetto, Laboratory of Ecohydrology, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland

One of the main goals of ECOPOTENTIAL is to improve the monitoring capabilities of ecosystems and ecosystem services in protected areas and natural ecosystems, providing trends and future projections under possible scenarios of climate change. The pillar idea to obtain such projections is to combine Earth Observations, in particular data derived from satellite remote sensing (SRS), into ecosystem models which numerically simulate the relevant dynamics of interest. In fact, EOs are a valuable source of information to characterize several input parameters and temporal forcings of ecosystem models, for example to better quantify the interaction between the vegetation component and the hydrological, energy and nutrient cycles. While this combination is a common practice for large-scale models (e.g. global or continental models), several difficulties may arise when SRS data is integrated to ecosystem models at the scales of protected areas. Limiting factors to local-scale applications are the temporal and spatial resolution of the data, the public unavailability of several remote sensing products, and the uncertainty associated to these measurements.

Vegetation at GPNP from Sentinel 2b

This review paper, written thanks to a collaborative effort among many ECOPOTENTIAL partners, provides examples of possible solutions adopted in recent literature and propose strategies to overcome scale issues and better quantify the model uncertainty introduced by SRS data. In particular we discussed three main integration possibilities: 1) SRS data as input to define model parameters and drivers, 2) SRS data as reference to validate model results, and 3) SRS data used within the context of data assimilation methods to improve the model predictions.

Sentinel 2b

Projects such as ECOPOTENTIAL, promoting the collaboration between the remote sensing and the ecosystem modeller communities, are essential to standardize techniques exploiting the full potentiality of SRS data for local applications.