Willow Flycatcher Habitat Model Results [ds278]

<DIV STYLE="text-align:Left;"><DIV><DIV><P><SPAN>This dataset was developed by Chris Stermer (CDFG - RAP Program). No original metadata were located, but the following is an abstract from a document describing the product: We conducted field surveys for Willow Flycatchers (Empidonax traillii brewsterii) in 1997 and 1998, from June 15 through July 31, within the McCloud Flats region of Siskiyou County, California. A Geographic Information System (GIS) was used to predict potentially suitable habitat to survey prior to field visits. We used a GIS to model willow flycatcher habitat within our study area from remotely sensed data and digitally mapped data layers. Spatially explicit data used in our predictions included a vegetation map (a vegetation classification derived from Landsat 5 Thematic Mapper imagery), a Digital Elevation Model (DEM), a slope gradient model, and a stream layer. Seventy-seven Willow Flycatcher territories were found during our surveys. Nine of the territories were located within a large montane meadow complex (Bigelow Meadows) known to have Willow Flycatchers, the remaining territories (68) were predicted using a GIS pattern analysis. We characterized vegetation within .07 ha circular plots centered on sixty-six territories located in 1997. Riparian thickets &gt; 2 m in height was the most abundant vegetation type, making up 53% of the vegetation within the plots. Twenty-one percent of the vegetation was a composite of live green grasses and forbs. A pattern based habitat predictability model was developed using the 66 territories located in the 1997 field season as image training sites. The model integrated two environmental variables found to have predictive capability: (1) composition of vegetation classes, and; (2) slope gradient. An accuracy assessment indicated the model was 94% correct when predicting suitable habitat greater than 6 ac. We concluded that Landsat Thematic Mapper imagery, when applied in conjunction with other landscape data, was an effective technique to identify suitable Willow Flycatcher habitat for our study area. Currently, this technique is being used by the California Department of Fish and Game to identify habitat throughout Northern California. This dataset was modified on May 17, 2005 by Eric Haney of CDFG - Information Services branch. Modifications included addition of a Site_ID Field, and fields representing UTM Northing and Easting coordinates (using NAD83 Datum). These fields were added to assist in an effort to field validate the dataset. Note that not all UTM coordinates are located within habitat polygons. Depending on the irregular shape of the polygons, some of the utm coordinates are located outside the boundaries. These coordinates are only to be used for coarse navigational purposes. While there is no publication date planned, Region 1 staff are working to validate the model results. </SPAN></P></DIV></DIV></DIV>

本数据集由Chris Stermer（CDFG - RAP项目组）开发。未检索到原始元数据，以下为一份介绍该数据集的文档摘要：1997年与1998年6月15日至7月31日期间，研究团队在加利福尼亚州锡斯基尤县麦克劳德平原区域开展了柳纹霸鹟（Empidonax traillii brewsterii）野外调查。团队借助地理信息系统（Geographic Information System, GIS）在野外踏勘前完成了潜在适宜调查栖息地的预测工作。基于遥感数据与数字化地图图层，研究团队在研究区域内构建了柳纹霸鹟栖息地模型。本次建模所用的空间显性数据包括：源自Landsat 5专题制图仪（Thematic Mapper）影像的植被分类图、数字高程模型（Digital Elevation Model, DEM）、坡度梯度模型以及水系图层。本次调查共记录到77处柳纹霸鹟领地，其中9处位于已知有该物种栖息的大型山地草甸复合体（比格洛草甸）内，剩余68处领地则通过GIS模式分析预测得到。研究团队对1997年记录的66处领地中心的0.07公顷圆形样地内的植被特征进行了表征：高度大于2米的河岸灌丛为最优势植被类型，占样地内植被总量的53%；21%的植被为活绿草类与杂类草构成的复合群落。研究以1997年野外季获取的66处领地作为影像训练样本，构建了基于模式的栖息地可预测性模型。该模型整合了两类具备预测能力的环境变量：(1) 植被类别组成；(2) 坡度梯度。精度评估结果显示，当预测面积大于6英亩的适宜栖息地时，模型准确率可达94%。研究团队得出结论：将Landsat专题制图仪影像与其他景观数据结合使用，可有效识别本研究区域内柳纹霸鹟的适宜栖息地。目前，加利福尼亚州鱼类与狩猎部门（CDFG）正采用该技术在北加州全境开展适宜栖息地识别工作。本数据集于2005年5月17日由CDFG信息服务部的Eric Haney完成修订，修订内容包括新增Site_ID字段，以及代表UTM北坐标与东坐标的字段（采用NAD83基准面）。新增上述字段旨在辅助开展数据集的野外验证工作。需注意，并非所有UTM坐标均位于栖息地多边形范围内，受多边形不规则形状影响，部分UTM坐标会落在栖息地边界之外，此类坐标仅可用于粗略导航。尽管暂无出版计划，但第一区域（Region 1）的工作人员正致力于验证该模型的预测结果。