Pasture Structure And Biomass Dynamic Analyzing System.

This thesis stems from research at UREP (Grassland Ecosystem Research Unit)at INRA (National Institute of Agronomic Research) in France. The aim of thisthesis is to set up a field-scale system which investigates and analyzes the spatialstructure and dynamics of managed, permanent grassland. Biologists will use thissystem to monitor the biomass and traits (main characteristics for species) of thepasture, to simulate agricultural management activity, to determine efficientagriculture strategy and to predict the future trends of the pasture under changingclimatic conditions.In order to carry out spatially-explicit analysis, the present vegetation growthmodel deals with a large plot of grassland and divides it into many small cells. Eachcell has the same area (0.1 m2; considered to be a bit-size area of vegetation for adomestic herbivore) but not necessarily same traits. These traits are grouped intopatches, which are the combination of many common species in a same cell. Fourcommon patches are described into the literature and are used to describe thegrassland plot. Each cell is divided in four compartments depending on theirdevelopmental stage: green vegetative, green reproductive, dead vegetative anddead reproductive. This model not only considers the permanent structure of thepasture but also the agriculture activities which are simulated as three types ofevents: cut, fertilization, and grazing by animals. These events occur according tothe management described by the biologist.To maintain and utilize this model, a graphic user interface which isprogrammed in JAVA is implemented. This GUI interface is to specify the input ofusers and parameters from files. Meanwhile, it provides the methods to set thestructure of the pasture. Finally, the output data is produced into the format CSV orNetCDF and displayed with a graphic chart. In the future, this model will beintegrated into CAPSIS (Computer-Aided Projection of Strategies In Silviculture)platform which simulates the forest growth and presents a better result display (3Ddistribution graph chart).