Benton Lake Refuge Shelterbelt Habitat Suitability and Bird Use Study

Benton Lake Refuge Shelterbelt Habitat Suitability and Bird Use Study Service Unit: Benton Lake NWR Reporting Office: Benton Lake Complex Species or group: Upland habitat – Shelterbelts 2007 Introduction In the recent past, planting shelterbelts was advocated in the Great Plains as a method of increasing species diversity (Schroeder 1986 and others), particularly bird diversity. Shelterbelts were planted on Benton Lake NWR initially in the 1960s, but have been maintained and expanded as recently as the 1990s. Since that time, many of the shelterbelt trees and shrubs have died, which may be the result of recent drought conditions. No attempt has been made to irrigate or replant the shelterbelts. However, during prescribed burning, the shelterbelts have been protected with firebreaks. Since the planting of shelterbelts on the refuge, additional research and understanding of the effects of woody vegetation in the Great Plains indicates that shelterbelts may have undesirable, negative effects on grassland bird species, in particular (see Bakker 2003 for review). Grassland species of concern at Benton Lake refuge include marbled godwits, chestnut-collared longspurs, Baird’s sparrow, grasshopper sparrow, lark bunting and Sprague’s pipit. In addition, most of the species of shrubs and trees planted on the refuge are not native, and maintenance of these would likely be in direct contradiction to the Service’s biological integrity policy (601 FW 3.10). In 2008, Benton Lake NWR will start its planning process to develop a 15 year management plan. Making management decisions about the shelterbelts on the refuge will be a key topic in this planning process. The key question will be whether or not to remove shelterbelts or maintain them. Having sufficient background data on shelterbelt condition, quality and bird use will inform future management of these shelterbelts. The initial inventory of shelterbelts described here will address four management questions (objectives): 1) What is the current condition of shelterbelts on the refuge in terms of their contribution to bird species diversity (assuming that was the original management goal)? 2) What bird species are currently using shelterbelts on the refuge and their relative relative abundance? 3) How important are shelterbelts at BNL to species of management concern? 4) Are shelterbelts having a negative impact on grassland birds on BLNWR? Methods In order to assess the current condition of shelterbelts on the refuge in terms of their contribution to bird species diversity we used a habitat suitability model described by Schroeder (1986), and later revised (Schroeder et al 1992). This model identified several characteristics related to the quality of shelterbelts for wildlife, particularly species richness of birds. Although the habitat suitability model focused on species richness, the authors speculated that the same variables defined in the model are positively correlated with wildlife productivity. Key variables in the revised Habitat Suitability Model that affect species richness in shelterbelts include shelterbelt size, number of snags/ha, average height of tallest row and foliage height diversity (MacArthur and MacArthur 1961). The area of each shelterbelt was measured by mapping the perimeter on foot using Trimble GPS unit (sub-meter accuracy). The perimeter was defined as the maximum exterior extent of the canopy of woody vegetation. Any breaks in shelterbelts greater than 50m were mapped as separate units. The 50m cut-off is somewhat arbitrary, but based in part on studies showing that this is an appropriate scale for the effect of trees on grassland songbirds (Bakker 2003). Snags (dbh>10cm) in each shelterbelt were simply counted. The 10cm cut-off is based on data on natural nest cavity selection by tree swallows, the cavity nesters most likely to be found on the refuge. We measured average tree height and foliar height diversity in the tallest row at 10 locations. We used a clinometer to measure the tree or shrub height at each point. Foliar height diversity was assigned at each point by indicating the number of strata (0-1m, 1-10m and >10m) in which vegetative layers reach their maximum height. The data for each variable corresponds to a suitability index variable (SI) (Schroeder 1992). The SI values are then entered in the following equation to get an overall Habitat Suitability Index value for the shelterbelt: HSIshelterbelt = SIfoliarheightdiversity x SIsnag x SIheight x SIsize 3 The HSI values for each shelterbelt are between 0 to 1, with 1 signifying an optimal shelterbelt for maximizing bird species diversity. Although not part of the HSI model, we also estimated the percentage of the shelterbelt that was dead and took several photographs. In order to identify what bird species are currently using shelterbelts on the refuge, and their relative abundance, each shelterbelt was surveyed four times between May 15 and June 16. One random, 100m x 40m belt transect per ha (approximately) was placed in each shelterbelt parallel to the nearest tree row (Quamen and Naugle, In press). We counted all birds seen or heard within 20m on either side of the transect, and noted bird movements to avoid double counting. We walked transects slowly (approximately 1 km/hr). All surveys occurred between sunrise and 11:00am and starting locations were rotated among all transects. Counts were not conducted on days with precipitation or excessive wind (>20km.hr; Ralph 1993). We used presence/absence and densities of birds within shelterbelts versus grassland areas as an indicator of negative impacts from shelterbelts. If grassland bird species occur in the grassland plots, but not the shelterbelt plots, or at higher densities in grassland plots, we will infer that these bird species are avoiding shelterbelts and the available grassland habitat for these passerine species on the refuge is reduced by the presence of shelterbelts. To do this, four grassland plots, with one 100mx40m transect each, were surveyed in the same manner and on the same days as the shelterbelt plots. Plots were selected randomly within native grasslands >800m from any shelterbelts (Quamen and Naugle, In press). The results of the four surveys of each shelterbelt and grassland transect were averaged to avoid pseudoreplication. Density differences between shelterbelts and grasslands were compared using t- tests. Each shelterbelt was completely searched for active nests twice during the breeding season (May 23-Jun 8 and Jun 18- Jul 6). For most nests, only the bird and tree species were recorded. However, for loggerhead shrikes, the primary species of concern that has historically nested in refuge shelterbelts, nests were checked every 3-5 days until the fate was determined. Nest and egg success were calculated based on exposure days (Mayfield 1975, Johnson 1979). Results The Habitat Suitability Index score for refuge shelterbelts ranged from a high score of 0.43 to a low of 0.09 on a scale from 0 to 1 (Figure 2). For most of the shelterbelts, less than 25% of the existing vegetation was dead (Figure 3). For details on each shelterbelts’ condition, size and photos see Appendix 1. Thirty-five bird species were detected in shelterbelt and grassland plots combined (Table 1). Twenty-four species were detected only in shelterbelts and two species, horned lark and clay-colored sparrow, were detected only in grasslands. Western meadowlark was the most commonly detected species (25%, n=104) in the shelterbelts and across all plots (23%, n=111). Chestnut- collared longspur was the most commonly detected bird in the grassland plots (34%, n=20). The habitat suitability model did a good job of relating shelterbelt condition to bird species richness for Benton Lake NWR (R2 =0.49, p=0.001). However, the shelterbelt next to headquarters, which scored the highest, had a strong influence on this relationship (Figure 4). We found a total of 44 active nests in the refuge shelterbelts. The most common species’ nest was Brewer’s blackbird, followed by eastern kingbirds and loggerhead shrikes (Table 2). Shelterbelt 12 had the highest number of nests (7). We found no active nests in six of the shelterbelts. The tree species most commonly used for nests was Russian olive (Table 3). Dead trees were used for 11% of the nests we found. We found 9 loggerhead shrike nests in 8 different shelterbelts. All nests successfully fledged young with an average of 5.78 ± 0.83 fledglings/nest. Shrikes nested in Russian olive (Elaeagnus angustifolia), both alive and dead, and caragana shrubs (Caragana arborescens). Swainson’s hawks are also a species of concern, but their nests were not monitored. They also nested in Russian olive and caragana. There were five species that were recorded on both shelterbelt and grassland transects and had at least 10 total detections. Of these, densities of chestnut- collared longspurs were significantly higher on grassland transects (Figure 4)(p=0.001). Densities of other species were not significantly different between the two habitat types (BHCO p=0.69, SAVS p=0.69, VESP p=0.23, WEME p=0.44). Discussion Based on the habitat suitability ranks for refuge shelterbelts, a few shelterbelts are in moderate condition and most shelterbelts are in poor condition relative to their potential to increase bird species diversity on the refuge. Shelterbelts ranked uniformly low for each of the key factors of size, structural diversity, snags and height. Most of these factors will not improve within shelterbelts unless there is active management to plant new trees/shrubs or irrigate to increase height, size and structural diversity. Such actions would be in direct violation of the Service’s biological integrity policy (601 FW 3.10). Some of the shelterbelts with caragana shrubs have expanded naturally, increasing size and structural diversity, which would likely continue slowly if left unmanaged. Russian olive does not appear to be expanding rapidly on the refuge as it can in certain areas (Tu 2003).

Load more