Estimating moist-soil seeds available to waterfowl with double sampling for stratification.

摘要

A study was conducted during autumns of 2001 and 2002 at Yazoo National Wildlife Refuge, west-central Mississippi, USA, to estimate the mean seed availability for 3 impoundments in each of 2 years and to determine if incomplete seed recovery from soil cores leads to biased estimates of seed availability. The statistical and cost efficiency of double sampling was compared to that of simple random sampling. The mean seed availability varied from 331-1084 kg/ha among impoundments and between years. With one exception, barnyard grass (Echinochloa crus-galli) and smartweeds (Polygonum pensylvanicum, P. lapathifolium and P. densiflorum) dominated all strata with mean seed availability of >=711 kg/ha. The exception occurred in the low density stratum of impoundment 4, where mud-plantain (Heteranthera reniformis) produced an unexpected abundance of small seeds. The precision of impoundment means, expressed as coefficients of variation, ranged from 7.0 to 11.5%, although most were <10%. Design effects for double sampling ranged from 0.44 to 1.02. Overall, 86.7% (672/775) of barnyard grass seeds were recovered and added to test cores. Percentages of seeds recovered from test cores did not vary with the number of seeds added, amount of detritus, or the interaction between these two factors. Using the null model, the estimated percentage of seeds recovered was 89.5%+or-2.2, and the ratio between seeds added and recovered (potential bias correction) was 1.123+or-0.027..