Sandbars managed for least terns within the Missouri River: evaluating the influence of fish, spatial scale and environment on habitat use

Abstract

Interior Least Terns (Sternula antillarum) nest on barren sandy habitats, typically sandbars and shorelines within large river systems of the central United States. Loss of natural ecosystem processes which create and maintain these habitats is considered the primary reason for population decline and Federal listing of this population. Throughout their range, management for Least Terns focuses on maintenance of breeding habitats, including placement of dredged material specifically as nesting substrate. Wide expanses of barren land are known to attract Least Terns, yet specific small scale habitat features are thought to trigger nest habitat selection. Yet as a plunge-diving piscivore, Least Terns also rely on ready access to appropriately sized slender-bodied fish: <52 mm total length for adults and <34 mm total length for young chicks. It remains unknown whether sandbar construction is a benefit or a detriment to forage abundance for Least Terns. Lastly, the relationship between availability of nesting habitats and required foraging habitats remains ambiguous. This information is needed to refine understanding of Least Tern habitat needs from a foraging ecology perspective, and contribute to knowledge and potential value of habitat restoration efforts. This study evaluates 1) the shallow water (<1.5 m) fish community near river and mechanically created emergent sandbars during three nesting seasons (2006-2008), 2) nest-scale habitat selection to determine if nest habitats differ between constructed and natural sandbars, and it evaluates the consequences of this selection on nest success, and 3) the potential associations between specific habitat features, at differing spatial scales, associated with airborne and foraging Least Terns. The research effort focused on the 95 kilometers of Gavins Point Reach of the Missouri River, between Yankton, SD and Ponca, NE during the Least Tern nesting season in 2006-2008. We sampled fish within 15-16 areas every two weeks from late May - July to document the relative abundance, species richness, and size classes of fish. Using systematic surveys on sandbars every 2-3 days, we detected and tracked 869 Least Tern nests until eggs hatched or failed, on constructed and natural sandbars in the Missouri River examining them for evidence of microhabitat selection at the nest and 3 m from nest. Least Terns successful foraging sites (N=416) were compared to a paired nearby random location to evaluate evidence of habitat selection during successful foraging in 2007-2008. We used systematic surveys every two weeks from late May - July in each year to identify Least Tern airborne (foraging or flying) locations within the river corridor (2006:966 sites, 2007:2940 sites, 2008:2003 sites), recording them using spot mapping. We modeled the probability that an observation (random and bird) as an airborne tern using logistic regression and habitat variables derived from remote sensing. Lastly, to refine our understanding of behavior in specifying habitat use, we modeled the probability of a location as a foraging site. Fish relative abundance was negatively related to depth. Catches were dominated by schooling species, including emerald shiner, sand shiner, spotfin shiner, and bigmouth buffalo. Significant inter-annual differences in relative abundance were observed, with generally increasing trends in intra-seasonal relative abundance of shiners and the smallest size classes of fish (<34 mm). Significant differences in the fish communities between the sandbar types were not detected in this study. These results suggest that mechanical sandbar habitats host comparable fish communities at similar levels of relative abundance. Among nests, significant differences were observed in substrates, amount of debris, and measures of vegetation between natural and created habitats, and between microhabitat at nests and the surrounding area. In general, Least Tern nest sites had coarser and larger substrate materials at the nest, more debris, and shorter and less vegetation compared to areas within the 3 m surrounding area. Nests in constructed habitats had a greater proportion of coarse substrates and less vegetation or debris than naturally created habitats. Observed nest success among constructed sandbars was 1.8x greater than that recorded among nests on natural sandbars, but a greater proportion of nests on natural bars were in or adjoining moist habitats where they were frequently destroyed. The best supported model predicting nest success on constructed and natural sandbars each included positive associations with percentage of pebble substrate, but inclusion of additional habitat predictors differed by sandbar type. Selection of microhabitat characteristics at the nest and vicinity constrains this species to barren water and wind scoured habitats. Historically, scoured habitats and the comparative abundance of debris within sites may have been a cue indicating safe habitats as river stage decreased. Water management regimes during this study favored survival of nests on higher elevation sandbars which included limited areas of wet substrates, typical of mechanically constructed sandbars. Among the micro-site habitat assessments for foraging birds, differences were not observed in water depth availability between natural or mechanically created habitats. Based on AIC relative importance scores, the variables water depth and microhabitat characteristics were most critical among micro-site characteristics of successful foraging locations; turbidity and larger scale aquatic features were comparatively of less importance. Within the greater landscape, Least Tern airborne locations within the river corridor were best explained by multi-scale logistic regressions; birds concentrated in areas with higher proportions of sandbar and wet sand habitats (200 m radius), while avoiding trees (50 m radius). When applied to a reserved sample of random locations, results suggest that in any given year, 40-97% of river corridor habitats were likely unsuitable for airborne terns. Furthermore, habitat use differed by behavior; foraging birds approached nearer to wet sand patches and in areas with less emergent sandbar habitat than flying birds, with some habitat relationships differing by year. Despite the surficial visual uniformity of many aquatic habitats on the Missouri River, successful foraging by Least Terns was associated with shallow, slack-water microhabitats adjacent to emergent sandbars and within areas with a greater proportion of sandbar habitat. This study demonstrates the importance of incorporating both spatial scale and behavior into ecological assessments of Least Tern foraging habitats.