Estimating Mean and Current Annual Increments of Stand Volume in a Regrowth Eucalypt Forest Using Historical Landsat Multi Spectral Scanner Imagery

Abstract

A key component in the management of the forest resource, be it for wood production, conservation or both, is accurate growth and yield information. The incorporation of remotely sensed imagery into growth and yield calculations can offer potential solutions to some of the problems of predicting forest growth rates. In particular, the repeat coverage of the satellite sensors such as Landsat Multi Spectral Scanner (MSS) data has allowed the development of long-term archives of the changing environmental and growth conditions of forests. This paper utilizes a temporal sequence of current annual increment (CAI) in stand volume and mean annual increment (MAI) data for a regrowth Eucalyptusforest on the tablelands of South-East New South Wales derived from dendrochronology and compares it to 15 historical Landsat MSS images over a 20 year period. CAI data was compared to the normalized difference vegetation index (NDVI) and the simple ratio (SR) derived from the MSS data. MAI data was also derived from the dendrochronology for 20 sites and compared to the integrated NDVI calculated from the satellite imagery. Results indicated close correspondence between the CAI data and the temporal changes in NDVI and SR for the forest stand and significant relationships (r2 = 0.60 to 0.74, p <0.01) between the integrated NDVI and MAI. The methodology demonstrates that remotely sensed data provides a retrospective method for estimating volume MAI of forest stands as well as a method for ongoing monitoring of CAI provided that atmospheric and geometric corrections are applied to the data.