Further growth of wind power development has elicited growing concern over bird/wind turbine interactions. The California Energy Commission has continued its efforts to investigate, understand, and, when needed, find solutions to this matter. This investigation is a continuation of our 1992 Altamont Pass Study, Wind Turbine Effects on Avian Activity, Habitat Use, and Mortality in Altamont Pass and Solano County Wind Resource Areas (Orloff and Flannery 1992). Both the 1992 study and the present study were funded by the Energy Commission.
This study had two objectives. The first was to further analyze mortality and observation data collected during our original 1992 study to: a) assess the potential effect of several additional turbine-specific and site-specific variables on the observed death rate, b) conduct a more detailed analysis of end turbine mortality, and c) assess whether data we collected on raptor perching could help explain observed mortality. The second objective was to collect and analyze new mortality data to attempt to corroborate some of our original findings. New mortality data were collected in 1994.
The results of different bivariate analyses suggest that several new variables were associated with mortality; raptor mortality was associated with turbines that had higher tip speeds, larger rotor diameters (i.e., larger rotor-swept area), variable-pitch blades, and that operated more frequently (percent time in operation). For reasons we discuss in text, however, we question the association of rotor diameter and blade pitch with raptor mortality. We believe that all four factors should be the focus of future studies to further test their potential association with mortality.
Our comparative analysis of mortality among the six turbine types studied indicated that mortality was significantly higher at lattice-tower turbines with horizontal cross arms (lattice/horizontal) than at all other turbine types combined, while mortality was significantly lower at guyed-pipe turbines that at all other turbine types combined. Mortality rates at the other turbine types were moderate. The principal features of lattice/horizontal turbines that may have contributed to their high mortality rate include a high percent of time in operation and high tip speed. We believe that the primary reason mortality was significantly lower at guyed-pipe turbines was because these turbines operated far less frequently than any other turbine type.
A multivariate analysis (multiple logistic analysis), that included new variables as well as some of the original variables, indicated that only turbine position in row (end vs. non-end) and turbine proximity to canyon were significantly associated with mortality. These two variables appear to be dominant over all other variables included in the analysis. Our analysis of new, 1994 mortality data corroborated our original findings that both position in row and proximity to canyon were significantly associated with mortality. Other recent studies have reported that mortality appears to be higher at end turbines as well (Hunt 1994, Winkelman 1992a).
There has been speculation recently about the possible role of perching in mortality. Birds may be injured or killed when they attempt to fly through spinning blades looking for a place to perch or when they leave a turbine once the blades have started spinning. In addition, by perching on turbines that are not operating, raptors may become habituated to turbines making them less caution around spinning blades. Out original observation data suggested that raptors seldom perched on operating turbines. However, a recent observation of a raptor that died attempted to land on an operating turbine, as well as another recent study in Altamont Pass that indicated that some raptors remained perched on operating turbines for extended periods, suggest that this could be a more common cause for mortality that previously suspected. Perching or attempting to perch on operating turbines may be rare events, but so are turbine-related mortalities; they may be rare but highly associated.
Our perching data may partly explain why some species appear to be more susceptible to collision than others. Red-tailed hawks and American kestrels exhibited a high perching frequency and we found in our original study that they were killed more often than would be expected from their abundance in the study area. In contrast, the turkey vulture? low perching frequency may help explain its lower relative mortality rate. However, perching frequency does not explain relative mortality for golden eagles or common ravens. Golden eagles rarely perched on turbines, but their relative mortality was high, while common ravens often perched, but their relative mortality was low.
"Perchability" of turbines may contribute to observed mortality for some turbines types. Perching frequencies were highest at lattice/horizontal (including windwall) turbines, followed closely by lattice-tower turbines with diagonal cross arms, guyed-pipe turbines, and vertical-axis turbines. Perching at tubular-tower turbines was considered lower. The low perching frequency (low perchability) at tubular-tower turbines may partly explain their relatively low mortality rate, while the high perching frequency (high perchability) at lattice/horizontal turbines may partly account for their significantly higher mortality rate. For the other three turbine types (lattice/diagonal, guyed-pipe, vertical-axis), however, perching does not appear to help explain mortality. The design and operation of these three turbine types may make perching birds less vulnerable to collision mortality.
Perching frequencies were higher at end turbines than at non-end turbines. This may be one reason the mortality rate was higher at end turbines. End turbines may provide better perch sites because they often provide a better view of a canyon where prey densities may be higher. This may also be why turbine proximity to a canyon was significantly associated with mortality.
Our data suggest that mortality at different turbine types may be caused by many interrelated factors. It is probably a combination and interplay of factors that determines observed mortality. We consider the analysis and results discussed in this report to be exploratory. The analysis was limited by an inherent lack of variability in some of the variables. We hope that this report will assist others by narrowing the scope of possible contributing factors and by providing a basis for future research and resolution of this issue.