Variability in multicomponent ecological systems: a test of new framework

点击翻译按钮获取中文摘要

Recently we discovered that if we measure variation in an ecological variables such as population size, productivity of a forest, frequency of pest a outbreak, species loss over TIME, we can relate it EXACTLY to variation of this variable over SPACE. More precisely, variance over time equals variance across space when sites vary independently, is predictably greater if they vary in concert and predictably less if their variation is consistently different between sites. The discovered relationship offers general PREDICTIVE and DIAGNOSTIC opportunities for assessing variation of ecological factors of interest and, importantly, mechanisms underlying this variation. Through theoretical analyses, experiments with laboratory ecosystems, natural aquatic microcosms, and crop yields we propose to test soundness and utility of this relationship. ***Most usefully, the model allows predicting temporal (often hard to get) from spatial variation (often available from surveys, public records, maps, or satellite images). Because variability of populations defines extinction risks, harvest levels, and management options; and variability of ecosystem services affects our health, recreation, and local economies, ability to assess it from spatial patterns alone is advantageous. Linking variability of environment to that of biological processes and species is my long-term goal. In the short term, we propose to conduct several tests of the model: ***(1) Its basic expectations - we will test hypotheses about the effects of habitat connectivity and environmental variation on the predictive power of the underlying equation via lab experiments; ***(2) That the patterns suggested by the model are important in nature. We will do so by manipulating connectivity and habitat variability in a system of small natural ponds; ***(3) The practical significance of findings by performing `proof-of-concept' exploration - here, we will apply the model to global food production. We will use UN data on the 15 most important crops of the world to determine if they can be predicted from spatial variation or heterogeneity in ecological attributes agroecosystems such as precipitation amount, habitat fragmentation, fertilizer use, local economic metrics, transport network, and others. Because departures from the `default' or null predictions of the model are diagnostic, the specific links between spatial variation of these attributes and variation of each crop will permit us to identify underlying mechanisms. This should help uncover vulnerabilities in regional (by continent or subcontinent) food production and its determinants. Such research could thus assist relevant international and national agencies in evaluating and adjusting their policies as well as forge more informed interactions with conservation goals and policies. **