Subarctic boreal forest albedo estimation using ENVISAT/ASAR and SPOT
Terhikki Manninen(1) and Aku Riihelä(1)
(1) Finnish Meteorological Institute, P.O. Box 503, FI-00101 Helsinki, Finland
The boreal zone land cover type has a very significant influence to northern hemisphere albedo and it is the main factor in northern hemisphere carbon budget. Boreal forest is also a sensitive indicator to changes in local and global climate. Tree species distribution reacts to changes in mean temperature and moisture conditions in long term; forest leaf area index and defoliation indicate stress factors in shorter time scale. Also the time of the phenological phase transitions are important indicators of global climatological processes.
Surface albedo is one of the Essential Climate Variables (ECV) defined in the Implementation Plan for the Global Observing System for Climate in Support of the UNFCCC. Surface albedo is also one of the ECVs that are largely dependent on satellite observations. Parameters describing, on one hand the land cover classes in general and, on the other hand the forest properties in detail, are related to the bidirectional reflection distribution function (BRDF) affecting the surface albedo, which is needed as input for climate change modelling. The BRDF of land cover is estimated for operational surface albedo products using optical satellite images and land use classification information. For vegetated land cover semi-empirical relationships depending on the leaf area index (LAI) are used. Recently a method for boreal forest LAI retrieval using the VV and HH polarization of ENVISAT/ASAR has been developed. Thus there is potential also for finding a direct relationship between BRDF and the microwave essence of land cover.
The NIR band albedo of subarctic boreal forest was recently found to be correlated with the VV/HH polarization ratio of ENVISAT/ASAR multiplied by the SPOT NIR band reflectance. Also, the NIR band albedo is significant in that it tends to dominate forest broadband albedo because of the high NIR reflectivity of vegetation. Since the test site of the ground measurements (FMI-ARC premises, north of the polar circle) was rather limited, it was considered important to test the method also in a larger area. For that purpose the MODIS black-sky NIR albedo product MOD43 was used as validation data set. The MODIS NIR albedo was compared with a NIR albedo derived using only the SPOT image reflectance values. Alternatively the MODIS NIR albedo was compared with the SPOT NIR reflectance multiplied by the VV/HH polarization ratio of ENVISAT ASAR.
First the MODIS, SPOT and ASAR data sets were co-registered. Then the backscattering values of the ASAR were averaged to the resolution of MODIS (500 m) using truncating mean (50%) and masking pixels of other land use classes than forest according to the CORINE land cover information. Likewise the SPOT reflectances were averaged to the MODIS resolution. The SPOT albedo was based on the NIR reflectance and existing BRDF formula parameterized with NDVI.
The SPOT/ASAR based NIR albedo correlated better with the MODIS NIR albedo than the NIR albedo derived using only the SPOT image. The reason is, that in the subarctic boreal forest NDVI is more related to the understory vegetation, such as lichen, than to the canopy, whereas the VV/HH backscattering ratio correlates well with the leaf area index of the canopy. Thus the conclusion is that the optical surface albedo products can benefit from the inclusion of microwave data.