The 1982 El Chichon and 1991 Mount Pinatubo volcanic eruptions produced large volumes of stratospheric aerosols that affected the environmental indices estimated from the Advanced Very High Resolution Radiometer (AVHRR) sensor. As a result the Normalized Difference Vegetation Index (NDVI) used to derive several vegetation health products exhibits a negative bias between approximately 20°N and 20°S during the post eruption years. In this work a statistical method based on Empirical Distribution Functions (EDF) and simple assumptions regarding the state of global vegetation is used to correct these biases. Results show that the statistical technique effectively reduces the biases allowing the construction of consistent time series from this historical dataset for climate studies.