Linking theory and data to advance knowledge on species distribution

生态学本质上是一门多元科学，因为物种之间以及环境之间以非随机的方式相互作用。生态的这种特殊性已经通过(元)群落理论进行了理论上的研究。然而，对于经验生态学来说，在非实验环境中检验理论上提出的想法是具有挑战性的，或者对于保护生物学家来说，将这些理论发展应用于实践一直是具有挑战性的，主要是因为没有统计模型明确地在理论和数据之间建立联系。我研究的长期目标是在生态学理论和数据之间建立更紧密的联系，通过开发专门设计的多元模型来促进我们对生态系统和保护实践的了解，从而有效地利用关于物种分布和相互作用的丰富的生物学数据。我和我的团队将专注于三个具体目标，横跨生态学理论、统计生态学和自然保护。(1)我将在理论上将地方和区域进程联系起来，并在调查数据中测试这一想法。(2)提出了多变量时空点过程模型(MSTPPM)，使其能够显式地考虑多个数据源和采样工作量。(3)利用最近发展的多元时空统计方法，我们将把(元)群落理论应用于濒危物种的保护。我将结合数学和统计方法来研究理论，并依靠模拟、经验系统和保护问题来测试和将理论应用于现实世界，从而达到我的目标。在目标1中，我将通过从理论上将局部(Lotka-Volterra)和区域(元种群)过程联系起来，重新定义实现的胡辛森生态位。我还将在元社区理论模型中包括环境不稳定(例如，气候变化造成的)。我将开发一个多元统计模型，明确设计用于在经验系统中测试这些理论发展。在目标2中，我将提出MSTPPM，它是为准时数据而设计的，以考虑抽样工作并整合来自调查数据的信息。在目标3中，我将提出在保护实践中使用多变量时空模型来解释生物相互作用的方法，以更好地保护和拯救魁北克黎塞留河濒危鱼类物种和留尼汪岛濒危海洋鸟类物种。它将提出考虑到生态问题而设计的多元统计方法，从而允许开展新的研究途径。它为生态学的理论和实证发展奠定了坚实的理论基础和统计学基础。最后，我提出的发展在生态内外都有广泛的应用。

Ecology is by its very nature a multivariate science because species interact with each other and the environment in non-random ways. This particularity of ecology has been investigated theoretically through (meta-)community theory. However, it has been challenging for empirical ecology to test the ideas proposed theoretically in a non-experimental setting or for conservation biologists to apply these theoretical developments in practice mainly because there are no statistical models that explicitly make the link between theory and data. The long-term objective of my research is to forge stronger links between ecological theory and data to advance our knowledge of ecological systems and the practice of conservation by developing multivariate model specifically designed to efficiently use the wealth of biological data being generated on species distributions and interactions. My group and I will focus on three specific objectives spanning ecological theory, statistical ecology and conservation. (1) I will theoretically relate local and regional processes across and test this idea on survey data. (2) I will advance multivariate spatiotemporal point process models (MSTPPM) for them to explicitly account for multiple data sources and sampling effort. (3) Using recently developed multivariate spatiotemporal statistical methods, we will apply (meta-)community theory to the conservation of an endangered species.  I will integrate mathematical and statistical approaches to investigate theory and rely on simulations, empirical systems and conservation problems to test and apply theory to real-world situations and thus reach my objectives.  In Objective 1, I will redefine the realized Huthinsonian niche by theoretically linking local (Lotka-Volterra) and regional (metapopulation) processes. I will also include environmental instability (e.g. resulting from climate change) in metacommunity theory models. I will develop a multivariate statistical model explicitly designed to test these theoretical developments in empirical systems.  In Objective 2, I will advance MSTPPMs, which are designed for punctual data, to account for sampling effort and to integrate information from survey data. In Objective 3, I will propose ways to account for biotic interaction in conservation practice using multivariate spatiotemporal models to better protect and save from extinction endangered fish species in the Richelieu River in Québec and endangered marine bird species on Réunion Island.  This program has multiple strength. It will propose multivariate statistical methodology designed with ecological problematic in mind, thus allowing to undertake new research avenues. It initiates solid foundations rooted in ecological theory and statistics to advance ecology both theoretically and empirically. Lastly, the developments I propose have broad applications within and outside ecology.