Stochastic dynamics models of plant population and species diversity

Although stochastic models of plant have been areas of much research in past few decades, people find it hard to choose an appropriate model among the many models available. This study provides answers to the questions of what and how plant population models should be used. Starting with the dynamics model and ending with the estimation model, this study discusses details of individual models and the connection of between those models. All the pieces are put together to provide the whole picture of the estimation of population dynamics.; Using the Shannon-Weaver index as an example, this study explains the problem of practical use of diversity indices. Reduction of some species due to disturbance can increase the species evenness, thereby increasing the value of the Shannon index. This is contrary to what is expected from disturbance. Therefore, one must look at the changes in other characteristics of the community such as the total population size and the distribution of species abundance.; Spatial distribution of species is very important to the ecological quality of natural systems. Two communities may have the same Shannon index value but their spatial distribution of species may be completely different. This study presents the spatial distribution of species diversity. The spatial variation of species is presented by a semivariogram. Ordinary kriging is used to interpolate the sample influence and the point Shannon index of all spatial locations.; This study presents sophisticated error analysis of the Shannon index in time and space. Error of current estimate increases proportional to measurement while the total forecasting error increases very fast in time. Spatial variation contributes the most to the total error of the Shannon index. Even under 100% measurement error, the spatial variation contributes over 65% to 80% while the measurement error contributes only 20% to 35%. Random error and modeling error contribute less than 5% to the total error.