Keywords
desert ecosystem; lizard; shrub; microclimate; machine learning
Start Date
26-6-2018 2:00 PM
End Date
26-6-2018 3:20 PM
Abstract
Deserts are characterized by extreme environmental conditions which require specialized adaptive mechanisms for both plants and animals to survive. Some of these important ecological mechanisms utilize indirect interactions involving non-consumptive synergy between species at distant trophic levels. For example, desert shrubs are known to mitigate extreme climate conditions and create a more favourable microclimate, thus providing thermal shelter for lizards. It is observed that changing climate may lead to more frequent and intense extreme weather events in deserts, such as prolonged drought and shorter intensive rainfall. It is, therefore, important to predict possible ecosystem dynamics under different scenarios of climatic change. In this study, we consider the interaction between a population of endangered blunt-nosed leopard lizards (Gambelia sila) and the dominant shrub Ephedra californica in a large desert ecosystem in the Panoche Hills, in San Joaquin Valley, California. On the basis of the observation data analysis, we conducted exploratory computations and applied machine learning techniques to build a predictive model of the microclimate conditions under the shrubs in a desert ecosystem, focusing on air temperature modification. It was demonstrated that the M5-Rules algorithm is able to generate predictive models that, on the unseen test subset, closely approximate observation data. Predictive models can be useful for the regional environmental authorities to plan for appropriate proactive measures.
An experience with data analysis and modeling of desert ecosystems
Deserts are characterized by extreme environmental conditions which require specialized adaptive mechanisms for both plants and animals to survive. Some of these important ecological mechanisms utilize indirect interactions involving non-consumptive synergy between species at distant trophic levels. For example, desert shrubs are known to mitigate extreme climate conditions and create a more favourable microclimate, thus providing thermal shelter for lizards. It is observed that changing climate may lead to more frequent and intense extreme weather events in deserts, such as prolonged drought and shorter intensive rainfall. It is, therefore, important to predict possible ecosystem dynamics under different scenarios of climatic change. In this study, we consider the interaction between a population of endangered blunt-nosed leopard lizards (Gambelia sila) and the dominant shrub Ephedra californica in a large desert ecosystem in the Panoche Hills, in San Joaquin Valley, California. On the basis of the observation data analysis, we conducted exploratory computations and applied machine learning techniques to build a predictive model of the microclimate conditions under the shrubs in a desert ecosystem, focusing on air temperature modification. It was demonstrated that the M5-Rules algorithm is able to generate predictive models that, on the unseen test subset, closely approximate observation data. Predictive models can be useful for the regional environmental authorities to plan for appropriate proactive measures.
Stream and Session
Stream B: (Big) Data Solutions for Planning, Management, and Operation and Environmental Systems
B2: Hybrid modelling and innovative data analysis for integrated environmental decision support