USGS - science for a changing world

Northern Prairie Wildlife Research Center

  Home About NPWRC Our Science Staff Employment Contacts Common Questions About the Site

Reproductive Success of Belding's Savannah Sparrows
in a Highly Fragmented Landscape

Results


We found a positive relationship (r² = 0.86, P < 0.01) between indices of sparrow population size and wetland size for the 26 marshes where Belding's Savannah Sparrows have been found over the past 20 years (U.S. Fish and Wildlife Service unpublished data; Fig. 2). The relationship may have been stronger, but was influenced by the small number (n = 4) of wetlands larger than 100 ha in southern California. Belding's Savannah Sparrows were found in all but four of the 26 salt marshes in 1986; all of the marshes where no sparrows were found were smaller than 10 ha (Zembal et al. 1988).

GIF -- Relationship between Belding's Savannah Sparrow populations and wetland size
Fig. 2. Relationship between Belding's Savannah Sparrow populations and wetland size. Population data are from a 1986 statewide census (Zembal et al. 1988).

By 23 March 1995, 54 males had established territories in the six plots, with the lowest densities found at F Street Marsh and the highest within the main Sweetwater Marsh complex (Table 1). Two males at D Street, one at F Street, and one at SM failed to attract a mate during the breeding season, whereas five males (SM, SH1, and SH2) were serial polygynists, attending two different females on each of their territories. In 1996, 45.5% of the banded males established territories on the same plots as in 1995. None of the three banded males at F Street, however, was seen in 1996. No birds banded as nestlings at Sweetwater Marsh were seen after fledging. Nesting densities (number of territories/study plot) were negatively correlated with mean territory size (r² = 0.68, n = 6, P = 0.04). Territory size ranged from 84.5 to 999.5 m²; no differences were found among the six plots (Table 1). None of the birds banded at Sweetwater Marsh prior to the 1995 breeding season was subsequently seen at any other marsh within the San Diego Bay area.

Table 1. Number, size ( SD), and reproductive index (based on ranking system of Vickery et al. [1992]) for Belding's Savannah Sparrow territories on 1-ha plots in three marsh types (centered in bold print) in coastal California in 1995.
  Reproductive index
Study Plot Habitat type No. territories Territory size (m2) Low (<5) High (≥5)
Large, connected
SH1 High-marsh 8 545.2 184.6 0 8
SH2 High-marsh 11 496.1 166.1 1 10
SM2 Mid-marsh 15 303.7 159.1 6 9
Vener Pond Mid-marsh 8 382.3 229.4 4 4
Small, Connected
D Street Mid-marsh 9 543.6 281.7 5 4
Small, Isolated
F Street Mid-marsh 3 625.5 186.9 3 0
All Sites 54 447.2 218.6 35% 65%

Based on the reproductive index, 65% of the breeding pairs received a rank of 5 or higher (Table 1). Comparisons of reproductive indices among plots (F = 7.9, df = 3 and 48, P < 0.01) showed differences between higher ranks within high-marsh plots (SH1 and SH2 combined), and lower ranks at F Street (P < 0.01), where no fledglings were seen. Likewise, comparisons among fragment categories (F2 = 5.6, df = 2 and 49, P < 0.01) showed that large marshes had higher reproductive indices (P < 0.05) than small, isolated marshes (Table 1). The number of fledglings observed per nesting pair over the breeding season ranged from 0 to 4 (GIF -- Mean x symbol = 1.2 SD of 1.2). This figure was an underestimation, however, because young Savannah Sparrows behave cryptically after leaving the nest (Ross 1980a), and few nestlings were banded. We found no relationship between territory size and reproductive success.

Eight of the 14 habitat variables entered into the forward stepwise DFA, with percent cover of bare ground and mean plant height being the best predictors of high-success versus low-success territories (Table 2). High-success territories were classified correctly in 88% of the cases using DFA, but low-success territories were classified correctly in only 37% of the cases. When we developed a model for territories versus non-territories, the best discriminators were percent cover of low-marsh plants, percent cover by dune species, and percent bare ground. The model correctly classified territories (64%) and non-territories (68%) based on 8 of the 14 habitat variables. Habitat characteristics of low-success territories were similar to those of high-success territories with respect to percent cover by bare ground and low-marsh plants, but overlapped with non-territories for plant height and percent cover by high-marsh species (Table 2). Few of the study plots, with the exception of F Street, had much space unoccupied by territories except for those areas with high percent cover by bare ground. Most of the vegetated areas consisting of mid- and high-marsh plant species were occupied by territories throughout the breeding season.

Table 2. Habitat variables ( SD) sampled at Belding's Savannah Sparrow territories (high and low success) and "non-territories" (unoccupied areas) in coastal California in 1995.
  Discriminant function analysisa
Variable High success
(n = 172)
Low success
(n = 82)
Non-territory
(n = 337)
Territory vs.
non-territory
High success
vs. low success
No. plant species 5.2 2.5 5.8 2.2 4.8 2.8
Max. height (cm) 56.8 18.2 49.6 17.1 45.7 25.6
Mean height (cm) 27.2 11.6 21.2 8.8 19.8 12.6 -0.17 -0.77
CV height (cm) 3.5 1.7 4.0 pm 1.3 3.6 2.1 -0.31
% Bare ground 13.4 26.5 15.5 25.5 25.8 34.8 0.41 -0.64
% Litter 11.7 23.4 16.8 28.9 9.9 20.3 -0.36 0.21
% Channel 3.1 11.2 4.6 17.8 2.6 12.1 -0.19 0.17
% Total plant cover 77.9 23.2 76.5 22.0 66.8 33.4
% High-marsh spp. 73.8 49.9 52.2 49.2 40.9 45.4 -0.22 -0.35
% Mid-marsh spp. 31.9 31.8 41.2 27.9 30.7 31.4
% Low-marsh spp. 8.8 16.7 18.8 20.5 20.9 25.9 0.69 0.27
% Edge spp. 2.8 8.1 0.4 0.2 1.1 5.6 -0.29
% Exotic grasses 0.4 0.3 0.5 0.3 1.1 5.3 0.24
% Dune spp. 0.0 0.1 0 1.3 5.9 0.47
aValues are correlation coefficients for forward stepwise discriminant function analysis.

Six of the seven nests we located successfully hatched young. Four of the seven nests were built within vegetation dominated by M. littoralis, and another nest was built in both M. littoralis and S. subterminalis. The remaining two nests were built within S. virginica and B. maritima. Nests were placed close to the ground ( = 3.3 ± 5.3 cm) where the amount of plant cover surrounding the nests was high ( = 99.0 ± 1.4%).


Previous Section -- Study Area and Methods
Return to Contents
Next Section -- Discussion

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://www.npwrc.usgs.gov/resource/birds/belding/results.htm
Page Contact Information: Webmaster
Page Last Modified: Friday, 01-Feb-2013 18:12:09 EST
Sioux Falls, SD [sdww54]