Topographic Metrics for Improved Mapping of Forested Wetlands

Topographic Metrics for Improved Mapping of Forested Wetlands
复制标题

用于改进森林湿地测绘的地形指标

DOI:
10.1007/s13157-012-0359-8
复制
发表时间:
2013
期刊:
影响因子:
2
通讯作者:
I. Yeo
I. Yeo
中科院分区:
环境科学与生态学4区
文献类型:
--
作者:
M. Lang;G. McCarty;R. Oesterling;I. Yeo

文献摘要

被引文献

相似文献

本文研究了基于光探测和测距(LiDAR)数据的数字高程模型(dem)和多种地形指标(包括多种地形湿度指数(TWI))生成的森林湿地地图的预测强度,其中TWI增强了出水口信息、归一化地形和TWI/地形混合地形)。将基于LiDAR DEM的湿地地图与淹没地图和现有湿地地图进行比较。基于最分布的FD8 (8 cell)和部分分布的D∞(1 - 2 cell)流路由算法的TWI与淹没的相关性优于基于非分布的D8 (1 cell)流路由算法的TWI,但D∞TWI类边界显得人为。增强后的FD8 TWI能较好地预测湿地位置,但不能预测淹没周期。归一化地形可以很好地预测洪水周期,但不能很好地绘制湿地边界。这些指标的混合提供了湿地位置和淹没周期的良好测量。基于地形指标的湿地地图包括与航空摄影湿地地图相似的洪水森林区域。这些结果表明,基于激光雷达的地形指标具有提高湿地制图精度和自动化的潜力。
We investigated the predictive strength of forested wetland maps produced using digital elevation models (DEMs) derived from Light Detection and Ranging (LiDAR) data and multiple topographic metrics, including multiple topographic wetness indices (TWIs), a TWI enhanced to incorporate information on water outlets, normalized relief, and hybrid TWI/relief in the Coastal Plain of Maryland. LiDAR DEM based wetland maps were compared to maps of inundation and existing wetland maps. TWIs based on the most distributed FD8 (8 cells) and somewhat distributed D∞ (1–2 cells) flow routing algorithms were better correlated with inundation than a TWI based on a non-distributed D8 (1 cell) flow routing algorithm, but D∞ TWI class boundaries appeared artificial. The enhanced FD8 TWI provided good prediction of wetland location but could not predict periodicity of inundation. Normalized relief provided good prediction of inundation periodicity but was less able to map wetland boundaries. A hybrid of these metrics provided good measurement of wetland location and inundation periodicity. Wetland maps based on topographic metrics included areas of flooded forest that were similar to an aerial photography based wetland map. These results indicate that LiDAR based topographic metrics have potential to improve accuracy and automation of wetland mapping.