Llwellyn M. Armstrong

Waterfowl Use of Dense Nesting Cover in the Canadian Parklands

Abstract Dense nesting cover (DNC) has been a conspicuous component of habitat management for upland-nesting ducks for >30 years, but its benefits for nesting ducks have been contentious. During 1994–1999 we monitored 3,058 dabbling duck (Anas spp.) nests in 84 DNC fields located throughout the Canadian Parklands to examine sources of among-field variation in nest density and nesting success. Nest density averaged 1.51 (SE = 0.15) nests/ha and overall nesting success was 20.4%, but there was pronounced annual variation in both estimates. Nesting success increased with increasing field size (range = 6–111 ha), but nest density remained constant. Nest density increased with percent wetland habitat within DNC fields and declined with percent perennial cover in the surrounding 2.4 × 2.4-km landscape, but these variables were not important for predicting nesting success. Nest abundance and nesting success roughly doubled in fields seeded with alfalfa (Medicago sativa) or sweet clovers (Melilotus spp.), but there was no benefit from using native as opposed to tame grasses. We recommend that waterfowl managers in the Canadian Parklands establish DNC with alfalfa in large fields in landscapes with abundant wetlands but minimal competing cover.

Population Dynamics of Breeding Mallards in the Great Lakes States

Abstract Mallard (Anas platyrhynchos) populations in the United States portion of the Great Lakes region increased through the 1990s but have since declined. To promote sustainable growth of this population, managers need to understand how perturbation of vital rates will affect annual population growth rate (λ). We developed a stage-based model representing the female mallard population in the Great Lakes using vital rates generated from a landscape-level study documenting reproductive parameters from 2001 to 2003. We conducted perturbation analyses (i.e., sensitivity analyses) to identify vital rates that most influence λ and variance decomposition analyses to determine the proportion of variation in λ explained by variation in each vital rate. Perturbation analyses indicated that λ was most sensitive to changes in nonbreeding survival, duckling survival, and nest success. Therefore, changes in these vital rates would be expected to result in the greatest Δλ. Process variation in breeding season parameters accounted for 63% of variation in λ. Breeding season parameters explaining the most variation were duckling survival (32%) and nest success (16%). Survival of adult females outside the breeding season accounted for 36% of variation in λ. Harvest derivation, high harvest, and high sensitivity of λ to nonbreeding survival for Great Lakes female mallards suggests there is a strong potential for managing the Great Lakes mallard population via harvest management. Because λ was highly sensitive to changes in duckling survival, we suggest programs that emphasize wetland protection, enhancement, and restoration as a management strategy to improve population growth for breeding mallards.

Estimating the nest-success rate and the number of nests initiated by radiomarked mallards

We developed a methodology for estimating the nest-success rate, the total number of nests initiated, and the average number of nests initiated per breeding pair by a group of radiomarked mallard (Anas platyrhynchos) females. Our methodology allows incomplete observation of all nests initiated and is related to current nest-success models. However, our method relaxes the assumption of equal daily survival rates made by the Mayfield method, and it incorporates the detection probabilities of newly encountered nests, which are ignored by the Mayfield method. Model selection and model averaging using Akaikes Information Criterion (AIC) are used so that estimates are no longer conditional upon the best-fitting model. Estimated daily nest-survival rates in our example were 0.94 (SE = 0.005), which gives an estimated nest-success rate over 25 days of 0.20 (SE = 0.03). We estimated that our sample of 124 resident females initiated 237 (SE = 15) nests for an estimated 1.91 (SE = 0.12) nests initiated per female. This was substantially larger than the observed (uncorrected) 1.41 nests per female, despite the high sampling effort and the fact that the nests were located at an average of 5 days of age. Our estimate of nests initiated was substantially more precise than similar estimates derived using the Mayfield method. We estimate an approximate 6-fold reduction in the sample size required by our method compared to the Mayfield method to obtain comparable precision for this parameter.

Long-term response of northern pintails to changes in wetlands and agriculture in the Canadian prairie pothole region

From 1955 through the late 1970s, northern pintail (Anas acuta) populations closely tracked the abundance of spring ponds. Declines in numbers of both northern pintails (hereafter, pintails) and ponds were evident during years of drought. However, since the early 1980s, the strength of the relationship between pintails and ponds has weakened greatly. Agricultural expansion on primary breeding grounds has been implicated as the cause of sustained pintail declines, but previous studies investigated pintail response only at large geographic scales (e.g., prairie-wide, stratum level). Potentially important effects of localized or multiscale changes in wetlands and agriculture on pintails are not well understood. Using data from the Canadian Prairie Pothole Region for 1961 to 1996, we investigated spatial and temporal covariation of pintail numbers with environmental factors (pond numbers and wetness indices) and agriculture at various scales. Models best supported by the data indicated that pintails responded positively to winter precipitation but with important regional variation and positively to pond numbers in some locations (southwestern Saskatchewan and southern Alberta). Results also indicated that pintail settling was better explained (increases in R 2 values of 0.05-0.06) using information about specific agricultural practices than about overall increases in farmed area. At a prairie-wide scale, we detected a negative association between settling and increased cropland area. At regional scales, settling was positively associated to various degrees with area in fallow (i.e., summerfallow-land tilled but not planted to crop in a given year). Both associations were strengthened with higher winter precipitation. Because cropland stubble is used readily as a nesting habitat by pintails and spring tillage of fields not used for summerfallow destroys nests, a shift from summerfallow to continuous cropping in the Prairie Pothole Region of Canada may have reduced the reproductive capacity of pintails in important breeding areas. In regions with characteristics that historically have attracted pintails to settle, we encourage land managers to promote agricultural practices that minimize use of spring tillage, convert cropland to perennial forages and pasture, and protect and restore wetland and upland habitat.

Waterfowl Nesting in Fall-Seeded and Spring-Seeded Cropland in Saskatchewan

Abstract Waterfowl nesting in annual croplands has remained a little-known aspect of waterfowl nesting ecology because of the inability of many studies to systematically search this habitat through the nesting season. Where searches have been conducted, they are generally restricted to the period prior to seeding, and many nests found are destroyed by the seeding operation. Consequently, fall-seeded crops have been promoted as an alternative cropping practice that could increase nest survival of waterfowl nesting in croplands. During 1996–1999, we conducted 3–4 complete nest searches on 4,274 ha of cropland, including spring-seeded wheat and barley, winter wheat, and fall rye in southern Saskatchewan, Canada. Using suites of predictive models, we tested hypotheses regarding relative nest abundance and nest survival among crop types and tested the influence of several landscape-scale covariates on these metrics. Apparent nest densities were higher in fall-seeded crops (winter wheat: 0.39 nests/ha, fall rye: 0.25 nests/ha) than in spring-seeded crops (0.03 nests/ha), and nest density in spring-seeded croplands increased with percent cropland and percent wetland habitat in the surrounding landscape. Nest survival was higher in winter wheat (38%) than in either fall rye (18%) or spring-seeded crops (12%), and nest survival in spring-seeded crops increased with relative nest initiation date. Nest survival was unaffected by surrounding landscape characteristics but tended to be higher in years of average wetness. Based on our findings, winter wheat and fall rye have the potential to provide productive nesting habitat for ≥7 species of upland nesting ducks and fall-seeded crops are a conservation tool well suited to highly cropped landscapes.

Impact of Management Treatments on Waterfowl use of Dense Nesting Cover in the Canadian Parklands

ABSTRACT Periodic treatment of established stands of dense nesting cover (DNC) is a recommended practice to maintain cover quality, but little information exists on the magnitude and duration of treatment effects on nesting waterfowl. During 1998–2001, we examined the effect of management treatments on vegetative characteristics and waterfowl nest success and density in fields of DNC seeded to introduced and native grass and forb mixes in the parklands of Saskatchewan and Manitoba. We measured vegetation height—density and litter depth within fields and located and monitored 1,927 duck nests within 33–42 fields/yr ranging in size from 6 ha to 62 ha. We considered a series of models examining the influence of grass type and management treatment (GTMT) and years post-management (YPM) on vegetative characteristics, nest success, and nest density while including covariates potentially affecting these response variables. Visual obstruction and litter depth were lowest in native-burned fields and greatest in introduced-hayed fields. Visual obstruction was low the year following management, peaked 2–3 YPM, and remained at intermediate levels through ≥6 YPM. Litter depth remained low for the first 3 YPM and increased thereafter. Nest success and nest density varied little among GTMT. Nest success was high (14.3%) the year following a management treatment, low (6.5%) at 2 YPM, and moderate thereafter. Nest success decreased with percent cropland in the surrounding landscape. Nest density was 0.7 nests/ha the first year following management, increased to approximately 1.3 nests/ha in years 2–3, and declined back to approximately 0.7 nests/ha for ≥6 YPM. Nest density decreased with field size and increased with the area of small wetlands, percent cropland, and percent wetland within surrounding landscapes. Nest density tracked vegetation density as expected and our results indicate a possible trade-off between nest density and nest success. Given ancillary data on small mammal and insect prey in our study fields, and evidence from other studies, we speculate that DNC fields may act as prey reservoirs during years of peak vegetative density with a consequent reduction in nest survival. Therefore, management to increase waterfowl production based on our results needs to consider the interaction of treatment effects, competing habitats, and surrounding landscape composition.

Mallard Nest-Site Selection in an Altered Environment: Predictions and Patterns

An animals reproduction and survival depend critically on the choices it makes when selecting a habitat for breeding. Therefore, we expect animals to select breeding habitats that confer fitness benefits. For mallards (Anas platyrhynchos), previous research has shown that populations are most sensitive to variation in nest survival rates; thus, we expect strong selection for safe nest sites. We used patterns of nesting success to predict nest distributions. We compared distributions of nests from radio-marked mallard females (n = 1,710) to random locations using logistic regression and modeled how nest-site locations were related to a number of landscape metrics measured at several spatial scales. Consistent with predictions derived from patterns of nesting success, mallard nest sites were best modeled using habitat attributes measured at the scale of the vegetation patch containing the nest, though ultimately habitat features measured at multiple scales improved predictions. Similarly, for most habitat ...

Wetland Selection by Mallard Broods in Canada's Prairie-Parklands

Abstract Although brood survival has a pronounced effect on population growth in mallards (Anas platyrhynchos), knowledge of brood ecology is more limited than for other vital rates. During 1993–1997 we collected wetland selection data from 210 radiomarked mallard broods on 15 study areas located throughout the Canadian Prairie-Parklands. We used information-theoretic approaches to select the best-approximating model of habitat selection in relation to wetland characteristics. Wetland permanence, cover type, width of flooded emergent vegetation, and interactions between these variables and date, moisture level, and dominant species of emergent vegetation were all important predictors of wetland selection. Mallard broods selected deeper wetlands, especially later in the brood-rearing season. Mallard broods also selected wetlands with large central expanses of open water and wide peripheral zones of flooded emergent cover. These habitat characteristics can most easily be met in landscapes that already contain an abundance and diversity of natural wetland habitats. Where such wetlands are unavailable, restoration or management of deeper wetlands may be necessary to meet the habitat requirements of mallard ducklings.

A Spatial Analysis of Factors Affecting Nesting Success of Shorebirds in the Canadian Prairies

Abstract. Conversion of natural grassland to cropland has been postulated as a cause for population declines among grassland birds. We evaluated nest-site habitat characteristics and landscape composition for three species of prairie-breeding shorebirds nesting across the agricultural landscape of prairie Canada. Nests of the Upland Sandpiper (Bartramia longicauda) and Willet (Tringa semipalmata) were found disproportionately in natural grazed (57%) or natural idled (21%) patches, whereas Marbled Godwit (Limosa fedoa) nests were in habitats in proportion to their availability on the landscape. We examined multiple landscape variables (proportions of crop, natural idled, natural grazed, and wetland habitats, edge density, and patch area) across a broad geographic scale, using relatively large samples of nests, and found no strong relationships. Neither nest-site habitat characteristics nor landscape-level effects influenced daily survival rates of Upland Sandpiper or Willet nests. Variation in landscape composition at three different spatial scales (200, 600, and 2000 m) also did not influence daily nest-survival rates. To adequately inform land-use management programs aimed at shorebird conservation, further research should examine site-selection patterns and characteristics of nest sites that maximize survival of chicks and adults during subsequent stages in reproduction.

Predicted Distribution and Characteristics of Wetlands Used by Mallard Pairs in Five Great Lakes States

ABSTRACT Understanding the relationship between wetland types and waterfowl distribution in the Great Lakes States of Wisconsin, Michigan, Illinois, Indiana and Ohio is complicated because basin specific waterfowl survey data do not exist. We used data from breeding waterfowl surveys in Michigan and Wisconsin during 1993 to 2002 and digital wetland data within buffered transect routes to develop a predictive model of mallard distribution within the 5 Great Lake States. The most parsimonious model based on AICc was used to map predictive distributions of breeding mallards. Based on the positive influence of palustrine emergent, palustrine unknown and palustrine unconsolidated shore and the negative influence of palustrine forested wetlands, the highest densities of breeding mallards were predicted in southeastern Wisconsin and southeastern Michigan. Additionally, we flew helicopter surveys in spring of 2003 to characterize wetland basins used by mallard pairs. Individual pairs were observed most often on small palustrine emergent and palustrine forested wetlands. The resulting models and maps can be used by a variety of agencies to plan conservation and management actions for mallards breeding in the Great Lakes States.