Search USGSSearch

Geology topics

Kathy M. Kraft

Publications and source records attributed to Kathy M. Kraft.

2 recordsLinked to original sources

Wetland use, settling patterns, and recruitment in mallards

The correlation between number of May ponds in the Prairie Pothole Region (PPR) of North America and size of the continental mallard (Anas platyrhynchos) breeding population the following spring weakened from the 1950s to the 1980s, suggesting possible changes in suitability of prairie ponds for meeting reproductive needs. We studied wetland use and preferences of radioequipped female mallards by reproductive stage (1988-90) in eastern North Dakota and westcentral Minnesota and evaluated effect of land use on pair distribution in eastern North Dakota (1987-91). May pond density varied among years and study areas, with changes in number of temporary and seasonal ponds accounting for 93% of variation in total ponds. During all reproductive stages, semipermanent basins were used most by females, but temporary and seasonal ponds were preferred during prenesting and egg production. Accounting for number of relocations, number of ponds used varied by year, by reproductive stage and with pond density during egg production. Numbers of breeding mallard pairs in stratum 46 in eastern North Dakota increased as May ponds increased from 1963 to 1985, but 33,659 fewer breeding pairs on average were present in 1971-85 than in 1963-70. Number of breeding pairs declined relative to May ponds from the 1960s to the 1980s, probably because fewer pairs settle in temporary and seasonal ponds as the percent of landscape in cropland increases. Waterfowl managers in the PPR should target efforts to increase duck production on landscapes where non-cropped temporarily and seasonally flooded wetland habitats are plentiful, thereby increasing cost effectiveness of management actions taken to increase nest success rate.

Journal of Wildlife Management

Using known populations of pronghorn to evaluate sampling plans and estimators

Although sampling plans and estimators of abundance have good theoretical properties, their performance in real situations is rarely assessed because true population sizes are unknown. We evaluated widely used sampling plans and estimators of population size on 3 known clustered distributions of pronghorn ( Antilocapra americana ). Our criteria were accuracy of the estimate, coverage of 95% confidence intervals, and cost. Sampling plans were combinations of sampling intensities (16, 33, and 50%), sample selection (simple random sampling without replacement, systematic sampling, and probability proportional to size sampling with replacement), and stratification. We paired sampling plans with suitable estimators (simple, ratio, and probability proportional to size). We used area of the sampling unit as the auxiliary variable for the ratio and probability proportional to size estimators. All estimators were nearly unbiased, but precision was generally low (overall mean coefficient of variation [CV] = 29). Coverage of 95% confidence intervals was only 89% because of the highly skewed distribution of the pronghorn counts and small sample sizes, especially with stratification. Stratification combined with accurate estimates of optimal stratum sample sizes increased precision, reducing the mean CV from 33 without stratification to 25 with stratification; costs increased 23%. Precise results ( x̄ CV = 13) but poor confidence interval coverage (83%) were obtained with simple and ratio estimators when the allocation scheme included all sampling units in the stratum containing most pronghorn. Although areas of the sampling units varied, ratio estimators and probability proportional to size sampling did not increase precision, possibly because of the clumped distribution of pronghorn. Managers should be cautious in using sampling plans and estimators to estimate abundance of aggregated populations.

North Dakota