Forest defoliation using IKONOS sensor for cork oak (Quercus suber L.) woods in Southern Spain

Abstract

Assessment of forest defoliation is severely hampered by the limited information on tree death on short temporal and broad spatial scales. In order to evaluate forest decline rates in cork oak of Southern Spain, an analysis was made of statistical correlation between the 4 IKONOS sensor original bands acquired in 2000, the Normalized Difference Vegetation Index (ndvi) and Infrared/Red Index (ir/r) and defoliation information obtained from the ground study. IKONOS near infrared band was negative correlated with defoliation (Pearson Correlation –0.762). The correlation between defoliation and ir/r ratio (Pearson Correlation –0.506) and ndvi (Pearson Correlation –0.449) was also significantly correlated. The dispersion of data presents in each category of defoliation justifies the use of the median value as the representative variable value for each intervals of defoliation. The Statistical Index of Defoliation (SID), generated from a lineal combination of IKONOS sensor bands, shows a correlation rate of 0.85 with the deforestation ground estimate which allows to estimate defoliation by using the equation: Y% defo calcu = –0.0016 * SID + 1.2162. This study shows that high-spatial-resolution satellite data can now be used to measure forest decline processes, suggesting many new alternatives to evaluate the impact of forest decline in Mediterranean forests.