Office Research and Development, National Health and Environmental Effects Research Laboratory, Western Ecology Division, U.S. Environmental Protection Agency, 200 SW 35th Street, Corvallis, 97333, OR, USA.
Biology Department, Kenyon College, Gambier, OH, USA.
Environ Monit Assess. 2019 Jun 20;191(Suppl 1):322. doi: 10.1007/s10661-019-7324-4.
In 2011, the US Environmental Protection Agency and its partners conducted the first National Wetland Condition Assessment at the continental-scale of the conterminous United States. A probability design for site selection was used to allow an unbiased assessment of wetland condition. We developed a vegetation multimetric index (VMMI) as a parsimonious biological indicator of ecological condition applicable to diverse wetland types at national and regional scales. Vegetation data (species presence and cover) were collected from 1138 sites that represented seven broad estuarine intertidal and inland wetland types. Using field collected data and plant species trait information, we developed 405 candidate metrics with potential for distinguishing least disturbed (reference) from most disturbed sites. Thirty-five of the metrics passed range, repeatability, and responsiveness screens and were considered as potential component metrics for the VMMI. A permutation approach was used to calculate thousands of randomly constructed potential national-scale VMMIs with 4, 6, 8, or 10 metrics. The best performing VMMI was identified based on limited redundancy among constituent metrics, sensitivity, repeatability, and precision. This final VMMI had four broadly applicable metrics (floristic quality index, relative importance of native species, richness of disturbance-tolerant species, and relative cover of native monocots). VMMI values and weights from the survey design for probability sites (n = 967) were used to estimate wetland area in good, fair, and poor condition, nationally and for each of 10 ecoregion by wetland type reporting groups. Strengths and limitations of the national VMMI for describing ecological condition are highlighted.
2011 年，美国环境保护署及其合作伙伴在美国大陆范围内首次进行了全国湿地状况评估。采用概率设计进行选址，以实现湿地状况的无偏评估。我们开发了一种植被多指标指数（VMMI），作为一种简洁的生物指标，适用于国家和地区尺度上的各种湿地类型。植被数据（物种存在和覆盖）是从代表七个广泛的河口潮间带和内陆湿地类型的 1138 个地点收集的。利用现场收集的数据和植物物种特征信息，我们开发了 405 个候选指标，这些指标具有区分受干扰最小（参考）和受干扰最大的地点的潜力。通过范围、可重复性和响应性筛选的 35 个指标被认为是 VMMI 的潜在组成指标。采用排列方法计算了数千个具有 4、6、8 或 10 个指标的随机构建的潜在全国范围的 VMMI。根据组成指标之间的有限冗余、敏感性、可重复性和精度，确定了性能最佳的 VMMI。最终的 VMMI 有四个广泛适用的指标（植物区系质量指数、乡土物种的相对重要性、耐干扰物种的丰富度和乡土单子叶植物的相对覆盖度）。使用概率点调查设计中的 VMMI 值和权重（n=967），估算全国和 10 个生态区按湿地类型报告组的湿地状况良好、中等和较差的面积。突出了全国 VMMI 描述生态状况的优势和局限性。
