Predicting invasion probabilities of introduced non-native freshwater fishes according to climate change and management

人为引起的气候变化导致气温变暖和降水模式的变化。虽然未来气温上升的规模取决于排放量，但升温高达2摄氏度被认为是不可避免的。全球变化的另一个方面是引入非本地物种（INNS）。虽然只有一小部分INNS发展为入侵，但对生物多样性的影响可能是不可逆转的。在温带地区，许多INNS建立，传播和影响（即入侵）的概率预计将在气候变化下增加。当预测表明优先行动的高风险物种时，INNS管理得到协助。INNS分布的预测基于物种分布模型，该模型基于非生物数据（例如气候）预测环境适应性。INNS的影响是通过经验和建模方法预测的。然而，这些方法不能整合改变的非生物和bioticconditions如何影响INNS和本地社区之间的紧急关系，或预测管理干预措施可以减少入侵probability.By相比之下，基于代理的模型（ABMs）明确考虑系统组件（代理），并试图了解系统属性如何出现从代理和他们的环境之间的相互作用。ABM是在不断变化的环境中以及在空间和监管背景下探索生态相互作用的日益强大的工具。然而，他们还没有被应用到了解气候变化和管理干预如何相互作用，影响INNS的入侵概率。因此，在英国的河流流域作为模型领域，淡水鱼作为模型INNS和环境署作为案例合作伙伴，博士开发新的ABMs预测当前和未来的INNS入侵概率根据气候变化预测，hydrologicalconnectivity和管理干预。ABMs将预测未来INNS的分布和英格兰河流域内部和之间的影响，以及一系列气候和管理情景。目标（O）将：O 1选择INNS模型，并生成当前和未来物种-气候关系、物种-物种相互作用的数据，并确定适当的管理干预措施;O2根据当前气候条件和水文连通性，以及人为干预措施，制定一个初始ABM，用于预测INNS在英国河流流域内的扩散和影响; O3对于每个模型INNS，参数化初始ABM以产生物种特异性最终ABM，以预测当前流域内和流域间的INNS入侵概率;和O 4使用最终的ABM来预测气候变化和管理干预将如何影响INNS在英格兰的未来入侵概率。在O 1中，水平扫描选择了三个在英格兰尚未入侵的模型INNS，它们代表了不同的生活史策略（强烈的r选择k）。模型参数化的数据，然后开发每个模型INNS使用现场，实验室和文献的方法。利用相关的空间、生态、水文和生物气候数据，在RANGESHIFTER中的O2中开发了初始ABM模型。最终的ABM是在O3中通过将O 1的参数和管理干预与O2的ABM整合而产生的。通过将气候变化预测与管理干预措施相结合，以产生与自然资源管理相关的知识，在O 4中预测性地使用它们。

Anthropogenic induced climate change is resulting in warming temperatures and changes inprecipitation patterns. Whilst the scale of future temperature increases is emissiondependent, warming of up to 2oC is considered inevitable. A further aspect of global changeis introduced non-native species (INNS). Whilst only a small proportion of INNS developinvasions, the impacts on biodiversity can be irreversible.In temperate areas, the probabilities of many INNS to establish, spread and impact (i.e.invade) are predicted to increase under climate change. INNS management is assisted whenpredictions indicate the high-risk species for priority actions. Predictions of INNSdistributions are based on species distribution models that predict environmental suitabilitybased on abiotic data (e.g. climate). INNS impacts are predicted by empirical and modellingmethods. These approaches cannot, however, integrate how altered abiotic and bioticconditions affect emergent relationships between INNS and native communities, or predicthow management interventions can reduce invasion probabilities.By contrast, Agent Based Models (ABMs) explicitly consider the system components (agents)and attempt to understand how system properties emerge from interactions betweenagents and their environment. ABMs are increasingly powerful tools for exploring ecologicalinteractions in changing environments and within spatial and regulatory contexts. They havenot, however, been applied to understanding how climate change and managementinterventions interact to affect the invasion probabilities of INNS.Thus, using river basins in England as model areas, freshwater fishes as model INNS and theEnvironment Agency as Case Partner, the PhD develops novel ABMs to predict current andfuture INNS invasion probabilities according to climate change projections, hydrologicalconnectivity and management interventions. The ABMs will predict future INNSdistributions and impacts within and between English river basins, and over a range ofclimate and management scenarios. Objectives (O) will:O1 Select model INNS and generate data on current and future species-climaterelationships, species-species interactions, and identify appropriate managementinterventions;O2 Develop an initial ABM for predicting the dispersal and impact of an INNS within Englishriver basins according to current climate conditions and hydrological connectivity, and nomanagement interventions;O3 For each model INNS, parameterise the initial ABM to produce species-specific finalABMs to predict current INNS invasion probabilities within and between river basins; andO4 Use the final ABMs to predict how climate change and management interventions willinfluence future invasion probabilities of INNS in England.In O1, horizon scanning selects three model INNS that are not yet invasive in England andrepresent different life history strategies (intense r-selected to k). Data for modelparameterisation are then developed for each model INNS using field, laboratory andliterature approaches. An initial ABM model is developed in O2 in RANGESHIFTER usingrelevant spatial, ecological, hydrological and bioclimatic data. The final ABMs are producedin O3 by integrating parameters and management interventions from O1 with the ABMfrom O2. They are used predictively in O4 by integrating climate change projections withmanagement interventions to generate knowledge relevant for natural resourcemanagement.