Author
CARVER, A - University Of Colorado | |
ROSS, J - Gm Sutton Avian Research Center | |
Augustine, David | |
SKAGEN, S - Us Geological Survey (USGS) | |
DWYER, A - Bird Conservancy Of The Rockies | |
TOMBACK, D - University Of Colorado | |
WUNDER, M - University Of Colorado |
Submitted to: Remote Sensing in Ecology and Conservation
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 2/6/2017 Publication Date: 3/27/2017 Citation: Carver, A.R., Ross, J.D., Augustine, D.J., Skagen, S., Dwyer, A.M., Tomback, D.F., Wunder, M.B. 2017. Weather radar data correlate to hail-induced mortality in grassland birds. Remote Sensing in Ecology and Conservation. doi:10.1002/rse2.41. Interpretive Summary: Large hail can kill animals, but its contribution to annual mortality is under-studied and difficult to measure. Hail events are hard to predict, and they often occur in locations where populations are not being studied. Animals such as songbirds are especially vulnerable to hail-induced mortality in rangelands where there is little vegetation cover to protect them. In this paper, we present a novel approach to estimating hail-induced mortality. We demonstrate a strong correlation between the distribution of nests destroyed by hail and that of Level-III Next Generation Radar (NEXRAD) data. A 22 June 2014 hail storm in northern Colorado destroyed 102 out of 203 known nests within our research site. NEXRAD data included atmospheric base reflectivity (BR), maximum estimated size of hail (MESH) and azimuthal wind shear (AWS), and these data were able to predict which portions of our study area received hail that was sufficiently large and intense to destroy songbird nests. Measurements of vegetation structure and cover at the nests sites did not correlate with hail-induced nest mortality, indicating that grassland vegetation provides insufficient cover to protect nests from large hail. We conclude that weather radar products can be used remotely to estimate hail mortality for grassland birds. Technical Abstract: Large hail can kill animals, but its contribution to annual mortality is under-studied and difficult to quantify. Hail events are challenging to predict, and they often occur in locations where populations are not being studied. Small-bodied terrestrial animals such as songbirds (Order Passeriformes) are especially vulnerable to hail-induced mortality. Focusing on the nest life stage in grassland songbirds, we suggest a novel approach to estimating hail-induced mortality. We demonstrate a strong relationship between the distribution of nests destroyed by hail and that of Level-III Next Generation Radar (NEXRAD) data. A 22 June 2014 hail storm in northern Colorado destroyed 102 out of 203 known nests within our research site, with a heterogeneous spatial distribution of mortality. NEXRAD data included atmospheric base reflectivity (BR), maximum estimated size of hail (MESH) and azimuthal wind shear (AWS). BR was strongly correlated with both MESH (CC=0.75) and AWS (CC=0.59), and MESH and AWS were also strongly correlated (CC=0.65); therefore we focused on MESH and AWS and did not consider the additive effect without the interaction term. For 133 ground nests where nest-site vegetation had been measured, we quantified the ameliorative influence of woody vegetation, nest cover, and vegetation density. We assessed the fit of 21 generalized linear models based on the independent and additive effects of weather and vegetation variables. A model including only MESH*AWS was most supported by the data (w=0.61), with less support (w=0.21, 'BIC'2) for models including vegetation or the independent effect of MESH or AWS. We conclude that weather radar products can be used remotely to estimate hail mortality among grassland birds, and that vegetation structure does not reduce vulnerability to hail for open-cup ground-nesting birds. We suggest the integration of remotely-sensed hail incidence in wildlife mortality estimation. |