Forecast future habitat suitability under climate change instructor materials
收藏资源简介：
In this module, learners apply a conservation lens to spatial modeling, taking on a case study of species distribution modeling around the Jack and Laura Dangermond Preserve in California, USA. In the problem-based lab, learners apply a MaxEnt machine learning approach to model habitat suitability for a species of concern (California red-legged frog) based on current climate conditions. Learners then forecast future habitat suitability under a range of future climate scenarios using the modeling parameters they developed. The module asks students to (1) acquire species occurrence data from the world’s largest open-source database of species observations, the Global Biodiversity Information Facility (GBIF), using a geoprocessing tool to query the GBIF API, (2) assess the data for quality, uncertainty, or bias, (3) work with multidimensional raster service layers for climate variables including temperature and precipitation, and (4) interpret model findings in the context of the spatial distribution of the data. Students will learn several analysis tools in ArcGIS Pro, including acquiring data from ArcGIS Living Atlas, working with multi-dimensional climate raster services, training a machine learning model with presence-only occurrence data, and interpreting model outputs and spatial map results. Links to the datasets used in the module are included in this Instructor Guide if instructors wish to have students apply these methods to another study area using the same data. Technical Requirements: Technical: This module uses ArcGIS Pro Modular: Learners will benefit from having knowledge of species observation data and climate datasets and the understanding of MaxEnt covered in the module Forecast species distinction under climate change. Key terms: GGA, analysis, cleaning, training, geoprocessing, source, API, distribution, map, database, climate datasets, GIS, case, are, platform, habitat, forecast species, spatial, predict species, data, raster, species, modeling, biodiversity, range, service, climate data, model habitat, species data, climate, index, area, model, context
本模块中，学习者将以保护视角开展空间建模实践，以美国加利福尼亚州杰克与劳拉·丹杰蒙德保护区（Jack and Laura Dangermond Preserve）周边的物种分布建模为案例展开研究。 在基于问题的实验环节中，学习者将运用MaxEnt机器学习方法，基于当前气候条件，为受关注物种——加州红腿蛙（California red-legged frog）构建栖息地适宜性模型。随后，学习者将利用自行构建的建模参数，针对一系列未来气候情景预测栖息地适宜性的变化。 本模块要求学习者完成以下四项任务：(1) 通过地理处理工具调用全球生物多样性信息设施（GBIF）的应用程序编程接口（Application Programming Interface, API），从全球规模最大的物种观测开源数据库——全球生物多样性信息设施（GBIF）中获取物种出现数据；(2) 评估数据的质量、不确定性与偏差；(3) 处理包含气温、降水等气候变量的多维栅格服务图层；(4) 结合数据的空间分布特征解读模型结果。 学习者将在ArcGIS Pro中掌握多款分析工具，包括从ArcGIS在线地图集（ArcGIS Living Atlas）获取数据、处理多维气候栅格服务、基于仅存在型观测数据训练机器学习模型，以及解读模型输出结果与空间制图成果。 若教师希望让学生使用相同数据集将本模块方法应用至其他研究区域，本教学指南中已附上本模块所用数据集的下载链接。 技术要求：本模块需使用ArcGIS Pro。前置知识要求：学习者需具备物种观测数据、气候数据集相关基础知识，以及对本模块《气候变化下的物种分布预测》（Forecast species distinction under climate change）中涵盖的MaxEnt方法的理解。 核心术语：GGA、分析、数据清洗、模型训练、地理处理、数据源、应用程序编程接口（API）、分布、地图、数据库、气候数据集、地理信息系统（Geographic Information System, GIS）、案例、平台、栖息地、物种预测、空间、物种预测、数据、栅格、物种、建模、生物多样性、范围、服务、气候数据、栖息地建模、物种数据、气候、指数、区域、模型、研究背景
