**3. Global biodiversity mapping efforts**

Early estimates (pre-2000) ranged from 4.3 to 5.3 million km2

**Figure 6.** Map of global waterbodies based on the Global Lakes and wetlands database (GLWD).

Open water constitutes almost the same spatial area, 422,111 km2

**Table 3** Global areal estimates of wetland coverages according to different studies.

**Study Wetlands (103**

Lehner and Doll [21] 9167 Williams [55] 8558 Mitch and Gosselink [56] 7000 - 9000 Mathews and Fung [50] 5260 Cogley [51] 4340 Sillwell-Soller et al. [52] 4795 GLCC [48] 1093 MODIS [49] 1291 Gross Wetlands Map [21] 11711 Finlayson et al. [57] 12800

timate inclusive of lake and reservoir waterbodies [57] relative to the reference [21] estimate

For the conterminous US, the Multi-Resolution Land Characteristics Consortium (MRLC) provides National Land Cover Databases (NLCD) as raster images [58]. According to the 2011 NLCD data, the area classified as woody or herbaceous wetlands sums to 417,442 km2

approach almost 13 million km2

Numbers provided by Lehner and Doll [21].

of 9.2 million km2

66 Pure and Applied Biogeography

whereas current estimates

. The USFWS maintains the

.

(**Table 3**). However, the highest estimate may be an overes-

. Within the US, wetlands are depicted by a few vector and raster datasets.

 **km2 )**

> Global and continental-scale mapping of freshwater species distributions has lagged freshwater habitat mapping efforts in terms of finer spatial granularity. More specifically, there are mismatches between the resolution of current global biogeography efforts and the spatial fidelity of waterbodies in the landscape. This makes intuitive sense for two main reasons: (1) The presence of a species within a given area typically requires in situ observation, as opposed to detection via remote sensing technologies, such as in the case of waterbodies and other landscape features. That being said, remote sensing of biodiversity is a rapidly growing area of research [61], with potential new capabilities for direct aerial observation of biota [62]. (2) Most observations of species are discrete points in space and time, are influenced by

methods of detection, and are not spatially comprehensive. Hence, extending species presences into unsampled areas requires various levels of inference ranging from summarization into regions or watersheds to sophistical statistical models predicting probability of presence using a suite of predictor variables characterizing habitat. Obviously, the first approach requires less resources and information, whereas the latter approach requires rich information on descriptions of habitat, not just the features themselves.
