Estimating Forage Production for Waterbirds and Waterbird Response to Habitat Management at Lacreek National Wildlife Refuge, South Dakota

Resource managers are often required to estimate waterfowl forage production on management units in order to meet habitat management goals. A technique was developed in Missouri to quickly estimate seed yields of common moist-soil plants consumed by a variety of waterfowl species by measuring phytomorphological measurements. Moist-soil seed prediction models based on similar phytomorphological measurements were developed for barnyardgrass (Echinochloa crusgalli), annual smartweed (Polygonum spp.), and beggarticks (Bidens spp.) on Lacreek National Wildlife Refuge, South Dakota in September of 2008 and 2009. Multiple linear regression was used to develop prediction equations for each of the target plant genera. Dry seed mass was regressed against various phytomorphological variables using Program R 2.10.1. Moist-soil plant models developed for Lacreek National Wildlife Refuge were compared to prediction models developed in Missouri and Mississippi to determine if models developed in southern moist-soil regions are applicable in northern moist-soil areas such as Lacreek National Wildlife Refuge. Models containing significant variables detected by studies in Missouri and Mississippi, but developed using my data, had lower predictability than my models for all plant genera common to each study. Southern models were also determined to predict significantly different seed yields than the Lacreek models and in some instances, southern models were significantly different than actual seed yields harvested at Lacreek National Wildlife Refuge. A technique was also developed to quickly estimate tuber and achene production of broad-leaved arrowhead (Sagittaria latifolia ) to be of value to waterfowl resource managers providing wintering habitat for trumpeter swans (Cygnus buccinator ). Tubers and achenes collected from arrowhead at Lacreek National Wildlife Refuge from September through October of 2008 and 2009 were used to create forage biomass prediction models based on phytomorphological measurements, leaf area, and environmental factors. Dry achene and tuber masses were regressed against various phytomorphological measurements, leaf area, water depth, and arrowhead canopy cover using Program R 2.10.1. Three models were developed to predict achene, tuber, and total forage biomass production of arrowhead. Coefficients of determination ranged from 0.63 for tubers to 0.83 for achenes. Duck use-days were used to calculate carrying waterfowl capacity on Lacreek National Wildlife Refuge based on average moist-soil plant seed and arrowhead forage yields. In 2008 and 2009, sampled impoundments produced an average of 150 kg/ha and 200 kg/ha, respectively. Using duck use-day estimates, it was determined that when moist-soil seed production on Lacreek National Wildlife Refuge was solely considered, the refuge is currently not capable of supporting the average number of dabbling ducks migrating and wintering on the refuge. In 2008 and 2009, sampled impoundments produced an average of 2,500 kg/ha and 2,000 kg/ha of achenes and tubers, respectively. Using waterfowl use-day estimates, it was determined that based solely on arrowhead tuber production, Lacreek National Wildlife Refuge is currently capable of supporting the average number of trumpeter swans wintering on the refuge and average arrowhead achene and tuber production is likely capable of supporting the average number of dabbling ducks migrating and wintering on the refuge. Population changes of secretive marsh birds were also analyzed to determine if moist-soil management practices were having any impacts on breeding marsh bird numbers. This study focused on habitat associated with breeding American bitterns (Botaurus lentiginosus), soras (Porzana carolina), and Virginia rails (Rallus limicola ) at Lacreek National Wildlife Refuge. Habitat feature measurements collected at use and non-use points included vegetation height, water depth, plant species composition, visual obstruction readings, and stem density counts. Based on habitat assessment measurements, logistic regression models, and comparisons using one-sided t-tests between habitat features at use and nonuse points, it was concluded that American bitterns were generalists as far as selection for habitat features within breeding territories was concerned. Soras tended to prefer breeding territories with 30% Typha spp. and 8% Carex spp. Virginia rails preferred vegetation heights ranging from 90-115cm in height and water depths ranging from 8-16cm with emergent wetland vegetation such as Typha spp., Scirpus acutus, and Sparganium eurycarpum interspersed with 10% open water. Results of my research provides guidelines for managers in terms of habitat manipulations that may be beneficial to breeding secretive marsh birds at Lacreek National Wildlife Refuge and may apply to other locations in the northern Great Plains.

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Thesis - University Access Only

Discipline thesis:degree_discipline

Wildlife and Fisheries Science

Year

2010

Author dc:creator

• McWilliams, Heather Nicole

Contributors dc:contributor

• Kent C. Jensen

• Natural Resources and Conservation

Statement dc:rights

• Copyright © 2010 Heather Nicole McWilliams. All rights reserved.

Language dc:language

en

Repository record dc:identifier

https://openprairie.sdstate.edu/etd/536

OAI identifier oai:identifier

oai:openprairie.sdstate.edu:etd-1536