Papers by Marc Vayssieres
Cross-channel variability in benthic habitat
Introduction Livestock grazing has been recognized as a major driving force for noxious weed inva... more Introduction Livestock grazing has been recognized as a major driving force for noxious weed invasions in pastures and on rangeland (Parker 1949). Livestock alter botanical composition and contribute to weed proliferation by reducing plant cover, dispersing seed, concentrating nutrients, compacting soil, and selective grazing (Burcham 1957). Many weeds that occur on grazing land possess anti-herbivore traits such as spines, stiff awns, high silica and lignin content, or secondary compounds such as alkaloids and glandular exudates. Because animals selectively graze some plants and avoid others, species that have grazing deterrents are favored on grazing lands and often increase relative to those eaten by livestock. Paradoxically, some noxious weeds that flourish on grazing lands have some stages of growth that are palatable to livestock, and with alterations in grazing management, can be suppressed by livestock. For example, medusahead (Taeniatherum caput-medusae) (Lusk et al. 1961),...
Yellow starthistle, a plant pest introduced to California in the mid7 8 0 0 ~ ~ has infested more... more Yellow starthistle, a plant pest introduced to California in the mid7 8 0 0 ~ ~ has infested more than 70 million acres and continues to spread. Vegetation managers, producers and land owners are searching for control methods that are compatible with their various land uses. Mowing and competitive plantings are two options that can be useful in yellow starthistle management programs. Timing is important. If mowing occurs too earlyy yellow starthistle can take advantage of the reduced competition for space, light and water. If it is done too late, large quantities of seed will disperse and replenish the seed bank. There were weed control benefits from planting subterranean clover as a competitive plant in combination with mowing, but the tested varieties declined substantially. Yellow starthistle is one of the most serious exotic plant pests in California. It was introduced in the mid-1800s and has now infested more than 10 million acres, according to the latest estimates. Yellow sta...
San Francisco Estuary and Watershed Science, 2010
An intelligent GIS for rangeland impact assessment
1993 4th Annual Conference on AI, Simulation and Planning in High Autonomy Systems, 1993
Journal of Range Management, 1999
We present a method for computerizing the transition rules of a state-and-transition model and th... more We present a method for computerizing the transition rules of a state-and-transition model and then linking this model to a geographic information system. The resulting simulation characterizes rangeland vegetation dynamics in space and time. The method makes use of an expert system, a computer program that forms logical chains of transition rules. Simulation using state-and-transition rules, sometimes called qualitative simulation, has the disadvantage that it is less precise than traditional numerical simulation. However, it may have the advantage of being able to generate more robust simulation of complex vegetation communities. We demonstrate the application of the method by constructing a model of hardwood rangeland in the western foothills of the Sierra Nevada. The model is tested by comparison with historic black-andwhite aerial photographs. The model is found to agree generally with the observed data but to differ substantially in some locations. Implications of this difference are discussed.
Journal of Vegetation Science, 2000
The use of Generalized Linear Models (GLM) in vegetation analysis has been advocated to accommoda... more The use of Generalized Linear Models (GLM) in vegetation analysis has been advocated to accommodate complex species response curves. This paper investigates the potential advantages of using classification and regression trees (CART), a recursive partitioning method that is free of distributional assumptions. We used multiple logistic regression (a form of GLM) and CART to predict the distribution of three major oak species in California. We compared two types of model: polynomial logistic regression models optimized to account for non-linearity and factor interactions, and simple CART-models. Each type of model was developed using learning data sets of 2085 and 410 sample cases, and assessed on test sets containing 2016 and 3691 cases respectively. The responses of the three species to environmental gradients were varied and often non-homogeneous or context dependent. We tested the methods for predictive accuracy: CART-models performed significantly better than our polynomial logistic regression models in four of the six cases considered, and as well in the two remaining cases. CART also showed a superior ability to detect factor interactions. Insight gained from CARTmodels then helped develop improved parametric models. Although the probabilistic form of logistic regression results is more adapted to test theories about species responses to environmental gradients, we found that CART-models are intuitive, easy to develop and interpret, and constitute a valuable tool for modeling species distributions.
Computers and Electronics in Agriculture, 2000
Ecosystem management, which is aimed at maintaining resources such as wildlife diversity, water q... more Ecosystem management, which is aimed at maintaining resources such as wildlife diversity, water quality, fire protection or aesthetic values, requires a regional or landscape perspective. Using a spatially explicit simulation model to test 'what-if' scenarios can facilitate the development of appropriate management practices at the landscape level. The state-and-transition model may provide a methodology that strikes an appropriate balance between simplicity and realism. We present a brief review of the state-and-transition methodology and then describe a computer-based implementation of a spatially explicit state-and-transition model. The implementation method is to link a rule-based qualitative simulation model with a geographic information system.
Journal of Vegetation …, 2000
The use of Generalized Linear Models (GLM) in vegetation analysis has been advocated to accommoda... more The use of Generalized Linear Models (GLM) in vegetation analysis has been advocated to accommodate complex species response curves. This paper investigates the potential advantages of using classification and regression trees (CART), a recursive partitioning method that is free of distributional assumptions. We used multiple logistic regression (a form of GLM) and CART to predict the distribution of three major oak species in California. We compared two types of model: polynomial logistic regression models optimized to account for non-linearity and factor interactions, and simple CART-models. Each type of model was developed using learning data sets of 2085 and 410 sample cases, and assessed on test sets containing 2016 and 3691 cases respectively. The responses of the three species to environmental gradients were varied and often non-homogeneous or context dependent. We tested the methods for predictive accuracy: CART-models performed significantly better than our polynomial logistic regression models in four of the six cases considered, and as well in the two remaining cases. CART also showed a superior ability to detect factor interactions. Insight gained from CARTmodels then helped develop improved parametric models. Although the probabilistic form of logistic regression results is more adapted to test theories about species responses to environmental gradients, we found that CART-models are intuitive, easy to develop and interpret, and constitute a valuable tool for modeling species distributions.
Uploads
Papers by Marc Vayssieres