Everglades Depth Estimation Network (EDEN) October 2011 Digital Elevation Model (DEM)

This is the 2nd release of the fourth version of an Everglades Depth Estimation Network (EDEN) digital elevation model (DEM) generated from certified airborne height finder (AHF) and airboat collected ground surface elevations for the Greater Everglades Region. Collectively, these
data are referred to as "High Accuracy Elevation Data" (HAED). Inconsistencies in available greater Everglades boundary files resulted in a missing elevation value for 1 cell along the western boundary of WCA3s. In this release, that single cell was "filled" with a value of
2.19682 meters as determined through the kriging process summarized below. As was the case with the 1st release (EDEN_em_OC11), this version differs from the previous elevation model (EDEN_EM_OCT07) in several ways. First, the kriging algorithm applied to newly modeled subareas was changed
from ordinary to universal kriging - resulting in slightly lower errors during cross-validation and accuracy assessment. Second, a previously omitted area in the northwestern corner of the Everglades National Park (ENP) and southern portion of Big Cypress National Preserve has been filled. Third, to increase accuracy in WCA1, the most challenging
EDEN subarea from an elevation modeling standpoint, the Conservation area is subdivided into 4 zones (Northern, Central, Southwest and Southeast). Boundaries between the North, Central and two Southern zones is based upon landscape units defined through the CERP. The landscape unit representing approximately the southern third of WCA1 was further divided into two zones (east and west) based on marked
changes in slope and aspect data generated from a DEM of the southern landscape unit as a whole. Division of WCA1 into 5 zones reduces errors estimated by comparing DEM modeled water depths with those measured by EDEN Principal Investigators in the field. Subdivision of the south landscape unit into east and west zones resulted in lower error estimates for the eastern subzone (i.e., southeast) without
significantly affecting (i.e., improving or degrading) the quality of the western subzone - an area where DEM modeling is most challenging. Finally, to reduce artificial breaks in elevation along WCA1 subarea boundaries, models were overlapped by 1 cell at these boundaries and, for the North, Central and South zone boundaries, overlapping model values were averaged. For the boundaries between the Southwest
and Southeast zones, cell values were "blended" based on weighted distance from the boundary edge. Finally, points along the North / Central and Central / South zone edges were subjectively selected and changed by adding or subtracting 0.03 meters (3 cm) to particular cells. This slightly reduces apparent artifacts without drastically
affecting the integrity of the model. The EDEN offers a consistent and documented dataset that can be used to guide large-scale field operations, to integrate hydrologic and ecological responses, and to support biological and ecological assessments that measure ecosystem responses to Comprehensive Everglades Restoration Plan. To produce historic and near-real time maps of water depths, the EDEN
requires a system-wide DEM of the ground surface.

本数据集为大沼泽地深度估算网络（Everglades Depth Estimation Network, EDEN）第四版的第二次发布，其数字高程模型（digital elevation model, DEM）基于经认证的机载测高仪（airborne height finder, AHF）数据以及空气船采集的大沼泽地区域地面高程生成，上述数据统称为‘高精度高程数据（High Accuracy Elevation Data, HAED）’。此前可用的大沼泽地边界文件存在不一致问题，导致WCA3s西部边界处有1个单元格缺失高程数值。本次发布中，通过下文所述克里金插值过程确定的2.19682米数值，对该单个缺失单元格进行了填充。与首次发布版本（EDEN_em_OC11）相同，本版相较于此前的高程模型（EDEN_EM_OCT07）存在多处优化。其一，针对新建模子区域所采用的克里金插值算法已从普通克里金变更为泛克里金，使得交叉验证与精度评估阶段的误差略有降低；其二，填补了大沼泽地国家公园（Everglades National Park, ENP）西北角以及大柏树国家保护区南部此前未覆盖的区域；其三，为提升WCA1（从高程建模视角而言难度最高的EDEN子区域）的建模精度，该保护区分设为4个分区：北部、中部、西南部与东南部。北、中及两个南部分区的边界依据《大沼泽地综合恢复计划》（Comprehensive Everglades Restoration Plan, CERP）划定的景观单元确定；占WCA1南部约三分之一的景观单元，则进一步基于从南部整体景观单元DEM生成的坡度与坡向数据的显著变化，划分为东、西两个子分区。将WCA1划分为5个分区后，通过对比DEM模拟水深与EDEN首席研究员野外实测水深所得的估算误差得以降低。将南部景观单元划分为东、西子分区后，东部子分区（即东南部）的误差估算值有所降低，且未对西部子分区的建模质量造成显著影响——西部子分区正是高程建模难度最高的区域。最后，为消除WCA1子分区边界处的高程人工断点，各模型在边界处重叠1个单元格；针对北、中与南部分区的边界，重叠单元格的数值取平均值。而西南部与东南部分区的边界，则依据距边界边缘的加权距离对单元格数值进行融合处理。此外，北/中以及中/南部分区边界的点位经主观选取后，通过对特定单元格加减0.03米（3厘米）进行微调，该操作可小幅弱化表观伪影，且未对模型整体完整性造成显著影响。EDEN提供了一套标准化且可溯源的数据集，可用于指导大规模野外作业、整合水文与生态响应，以及支撑针对《大沼泽地综合恢复计划》生态系统响应的生物学与生态学评估。为生成水深的历史与近实时地图，EDEN需要一套覆盖全系统的地面高程数字模型。