MAMU Collision Risk Assessment for 2018 & 2019
Total Page:16
File Type:pdf, Size:1020Kb
Marbled Murrelet Collision Risk Assessment Associated with the Humboldt Wind Project Proposed for Humboldt County, California: 2-Year Report Project #3980 Prepared for: Humboldt Wind, LLC 11455 El Camino Real, Suite 160 San Diego, CA 92130 Prepared by: H. T. Harvey & Associates Richard Golightly, Ph.D., Senior Associate Ecologist Stephanie Schneider, M.S., Wildlife Ecologist Scott B. Terrill, Ph.D., Principal September 2019 983 University Avenue, Building D Los Gatos, CA 95032 Ph: 408.458.3200 F: 408.458.3210 Executive Summary The marbled murrelet (Brachyramphus marmoratus) is a federally threatened and state endangered seabird that flies between the ocean and inland sites used for nesting. In southern Humboldt County, development of a wind energy facility has been proposed by Humboldt Wind, LLC., which would require construction of a line (or “string”) of turbines along the top of Bear River Ridge and Monument Ridge. There is potential for murrelets to encounter the wind turbines associated with this Project if they cross these ridges as they transit to and from nesting habitat. Marbled murrelet activity at the site of this Project was assessed using radar surveys conducted during the 2018 and 2019 murrelet breeding seasons (April through September). We used ridge passage rates determined by radar to develop two Collision Risk Models (CRMs) that assess potential collision risk for marbled murrelets: a deterministic model that provided a single conservative estimate of risk each year and a probabilistic model that quantified the probability of all possible estimates of risk which facilitated expression of uncertainty around the most likely estimate and produced an upper limit of potential risk each year. For a CRM to accurately estimate collision risk, it is essential to quantitatively assess variability in passage rate across time and space. For murrelets, inland flights are non-random and vary in predictable ways seasonally: with the majority of flights occurring during the breeding season, and topographically, with primary flyways being valleys rather than ridgelines. To appropriately capture this variability in inland activity, radar-based observations of murrelet flights were made across the length of the turbine string for the duration of the breeding season at least once each month. These radar surveys provided an accurate estimate of passage rates (birds/day/km) for the breeding period at monthly intervals. Passage rates were calculated for each month from the radar surveys. Calculating passage during the non-breeding period (October-March) was complicated because observations were extremely rare and required some extrapolation between neighboring months. Spatially, the turbine strings were divided into 11 zones based on large gaps (>800 m) between adjacent turbines and topographical differences (peaks, low points, and elevational drops into the adjacent river valleys). Both CRMs used passage rate, murrelet size, flight speed, and turbine characteristics to calculate the probability that individuals crossing the turbine string would pass through a rotor swept area of a turbine and, upon passage, collide with a rotor blade. To the maximum extent possible, input data was based on empirically derived values to ensure that estimates reported here were as realistic as possible. Flight speed, avoidance, and interannual variability in passage rate were each identified as having a range of possible values. For the deterministic model we assumed an average flight speed, a conservative value of avoidance (the probability that a bird will detect and avoid a turbine), and the best-fit relationship between inland activity and a reliable predictor of inland activity (ocean productivity). For the probabilistic model, we did not make these assumptions; rather, all possible flight speeds, avoidance rates, and interannual variability in passage rate (as informed by oceanic conditions) were considered using a Bayesian framework. Probabilities of occurrence were calculated for each potential value of these three variables which enabled estimation of collision risk while accounting for inherent variability and uncertainty needed to produce an upper estimate of risk. Marbled Murrelet Collision Risk Assessment H. T. Harvey & Associates i Humboldt Wind Project September 2019 Because turbine types have not yet been finalized by the Project, we assumed a worst-case scenario in terms of specifying the maximum turbine size and rotational speed in the CRMs. These assumptions were unrealistic and resulted in an overestimation of collision risk. In this maximum turbine size and speed scenario, the deterministic model estimated that 0.175 birds per year in 2018 and 0.246 birds in 2019 would collide with turbines associated with the proposed facility; if extrapolated over the 30-year life of the Project, there would be potential for 5.26 and 7.38 murrelet collisions (approximately 1 individual every 3 to 4 years) based on 2018 and 2019 data, respectively. When considering long-term (30- year) variability in ocean conditions associated with inland activity of murrelets, the number of collisions estimated over a 30-year period was 5.27 or 6.85 birds based on 2018 or 2019 data, respectively. Considering the year-specific outcomes of the deterministic CRM, this would be an average of 6.1 birds during the 30 years life of the Project. Like other deterministic CRMs, the model presented here was sensitive to change in avoidance probability. The deterministic CRM was also sensitive to changes in the annual passage rate of murrelets as well as operational factors specific to the turbines (e.g., blade speed) that were assumed to be maximal for estimation of collision risk. A benefit of the deterministic model was that differences in passage rates could be assessed along the length of the turbine strings. In 2018 the lowest estimate of passage rate among the 11 zones was 0.00 birds/km/day and the greatest estimate was 0.683 birds/day/km (푥̅ = 0.2136 birds/day/km for the entire length of the turbine string). In 2019 the lowest estimate of passage rate among the 11 zones was 0.078 birds/km/day and the greatest estimate was 0.973 birds/day/km (푥̅ = 0.2871 birds/day/km for the entire length of the turbine string). Generally, passage rate was greatest for zones that descended into adjacent river valleys, which is logical given that low-elevation valleys serve as primary travel corridors for inland flying murrelets. The deterministic CRM in 2018 was used to identify the three highest risk zones and all turbines were removed from those zones; this resulted in a 36% decrease in overall risk from the original 60-turbine Project layout. The updated 47-turbine layout was used for all results presented in this Collision Risk Assessment Report. The probabilistic model estimated an upper bound of 6.17 collisions over the 30-year life of the Project based on data collected in 2018 and 7.77 collisions over the 30-year life of the Project based on data collected in 2019; this upper bound represents the one-tailed threshold that contains 95% of the probability distribution. Thus, the maximum take given uncertainty in the measurements is estimated to be 7.77 birds over 30 years, with only a 5% chance of being greater. Like the deterministic CRM, these estimates represent a worst-case scenario in that the calculations assume operational factors specific to turbines (e.g., turbine size and blade speed) were maximal. Estimates produced by the probabilistic CRM corroborate estimates produced by the deterministic CRM. The benefit of a probabilistic approach was that it allowed for accounting of uncertainty and variability associated with model inputs, particularly those that the deterministic model was most sensitive to (specifically avoidance probability, interannual variability in passage rate, and flight speed). Although flight speeds were well documented, this input was relatively variable and accounting for all potential flight speeds provided a more realistic assessment of the range of risk rather than basing the assessment on the average flight speed. Unlike Marbled Murrelet Collision Risk Assessment H. T. Harvey & Associates ii Humboldt Wind Project September 2019 the deterministic CRM, collision risk was not able to be estimated by zone in the probabilistic model. Since both CRMs ultimately produced similar estimates of risk, we also assumed the deterministic model provided a reliable assessment of zone-specific risk. The results presented in this report are consistent with earlier reports of collision risk and consistent with conclusions of significance and associated mitigation measures provided within the Draft EIR for the Project. Marbled Murrelet Collision Risk Assessment H. T. Harvey & Associates iii Humboldt Wind Project September 2019 Table of Contents Executive Summary ............................................................................................................................................................. i Tables .............................................................................................................................................................................. vi Figures .......................................................................................................................................................................... viii List of Preparers ...........................................................................................................................................................