Species distribution modeling in regions of high need and limited data: waterfowl of China

Table 2 Validation measures testing errors of omission for China waterfowl distribution maps at three scales: 1, 5, and 10 km

Common species name	Scientific name	Omission error rate^a			Number of validation points^b
Common species name	Scientific name	1 km	5 km	10 km	Number of validation points^b
Bar-headed Goose	Anser indicus
Breeding season		0	0	0	13
Wintering season		0.095	0.095	0.095	21
Common Goldeneye	Bucephala clangula	0	0	0	74
Common Merganser	Mergus merganser	0	0	0	18
Common Pochard	Aythya ferina	0	0	0	1
Common Shelduck	Tadorna tadorna	0	0	0	2
Common Teal	Anas crecca	0	0	0	7
Greater Scaup	Aythya marila	0	0	0	5
Greylag Goose	Anser anser	0	0	0	1
Long-tailed Duck	Clangula hyemalis	0	0	0	40
Mallard	Anas platyrhyrnchos	0	0	0	166
Northern Pintail	Anas acuta	0	0	0	1
Red-breasted Merganser	Mergus serrator	0	0	0	64
Smew	Mergellus albellus	0	0	0	2
Tufted Duck	Aythya fuligala	0	0	0	8
Whooper Swan	Cygnus cygnus	0	0	0	17
Total number of validation points					406

^aOmission rate is calculated by dividing the number of correctly predicted presence locations by the total number of validation (presence) points. For example, of the 21 validation locations where Bar-headed Geese were observed during the wintering season, two (or 9.5%) were incorrectly predicted as “absent” within the grid cell (within 1 km). In this example, increasing the number of neighboring cells to 5 or 10 km did not improve the error rate
^bAsian Waterbird Census data were used as validation points for all species listed. In addition, for the Bar-headed Goose we tested for errors of omission using location data (both the breeding and wintering seasons) from our satellite telemetry work (Prosser et al. 2011)

ISSN: 2053-7166