Get reliable science

The identification and development of methodologies to facilitate sustainable rangeland management are essential for management of ecosystems supporting livestock production. Remote sensing could be an inexpensive method to determine forage production over large areas. However, most studies using this technology for biomass determination in rangelands were conducted in flat areas with uniform canopy. Using a multispectral sensor, we attempted to develop a regression model for determination of standing crop biomass in a dry scrubland in Mexico. The study was conducted from April 2009 to July 2010 at two field sites supporting dry-scrubland vegetation. The sites had different topographic and soil features, but similar vegetation types. Each site covered 225 hectares with 153 samples-plot. We collected samples four times throughout the year (two in the dry season, two in the rainy season). A multispectral sample nd a vegetation clip sample were taken in each plot. A radiometer multispectral sensor was located on a platform 3.2 meters above the soil surface. The sample reading covered an area of 1.6 m, which was the same area used for clipped sample. The reflectance values were used to derive an index of the reflectance (NDVI). Based on the NDVI and clipped biomass data, regression models were developed to predict the biomass production for the two seasons (dry and rainy). The coefficient of determination showed a high variability, varying between 0.14 in the dry season and 0.30 in the rainy season. The variability in the canopy height in shrub vegetation is a limiting factor to consider when using remote sensing. It was concluded that reflectance was not a good estimator of the amount of forage in dry scrubland. Consequently, improvements should be made before optical techniques are used in this type of studies.