Spatial and seasonal dynamics of rangeland herbage: An integration of proxy and direct monitoring approaches .

Abstract

Rangeland forage quantity and quality are subject to spatial and temporal variability mainly due to the inherent variations in rainfall and landscape characteristics. Consequently, regular forage condition assessment and monitoring to provide reliable information for grazing management are vital. Studies on herbage assessment and monitoring have mainly focused on quadrat based harvesting approaches which are limited to small areas and often based on a few samples. Remote Sensing (RS) and Geographic Information Systems (GIS) have been proved to be very useful technologies for rangeland forage monitoring. This study was designed to contribute to the development of an integrated spatial and temporal grazing management information system for assessing, monitoring and predicting rangeland forage quantity and quality by taking advantage of RS and GIS. In particular, the study investigated the spatial and seasonal patterns of herbage quantity and quality in relation to grazing, vegetation cover and soil type by integrating both RS and field data using GIS. Two approaches were used to assess, monitor and predict herbage quantity and quality: Proxy and direct harvest methods. Proxy methods of assessing quantity included extraction of vegetation physiognomic cover classification from satellite images and measurement of herbage cover and height in different vegetation types, which served as a basis for predicting herbage mass. Analysis of species composition was used as an indirect way of assessing quality in different vegetation strata. The direct methods included clipping, drying at 60oC, weighing and analysis of neutral detergent fibre, digestibility and crude protein content of herbage from 1x1m sample plots using standard methods. The effect of season, vegetation cover types, grazing, soil types and their interactions on herbage mass, nutritive value, species cover and height were analysed using Analysis of Variance (ANOVA). Herbage species cover significantly differed (p<0.05) across seasons with the highest herbage cover (77%) occurring during March-May wet season and the lowest (27%) during September-November wet season. Herbage species height ranged from 11 to 15 and 16 to 23 cm during the dry and rainy seasons respectively. Vegetation cover, soil type and grazing explained 85% of herbage dry matter yield, 77% of crude protein, 67% organic matter digestibility, and 64% neutral detergent fibre variations. Spatial variation of herbage was mainly influenced by grazing and vegetation cover. Ungrazed sites were 42% lower than grazed sites. Herbage yield on grassland patches was 21% higher than the yield from woodland. Results of vegetation classification from both Landsat and IKONOS images showed that grassland patches were classified more accurately compared to woodland patches. Grouping detailed vegetation classes to a definition level that creates a favourable relationship between sensor resolution and vegetation patchiness increased herbage mapping accuracy for both classifier and imagery type. This study demonstrated that herbage cover is an important proxy measurement of spatial and seasonal patterns of herbage mass. Results showed that vegetation cover type and grazing were the key factors in determining herbage species composition and quality. Use of fuzzy classifiers improved mapping accuracy in comparison to maximum likelihood classifiers. In the quest to further improve rangeland herbage mapping, there is need to investigate other classifiers. . It has been demonstrated that herbage quantity and quality can in reality be monitored based on cover and species composition measurements to avoid or at least minimise the cost, destruction and information timeliness implications that are known to be associated with harvesting methods. Results from this study also showed that grazing and vegetation cover management are essential for rangeland productivity and biodiversity conservation. The substantial changes in temporal patterns of herbage composition resonate the need for regular monitoring and provision of information for sustainable rangeland ecosystem management