AUDUBON CANYON RANCH RAVEN PROJECT PROGRESS REPORT, DECEMBER 2001

AUDUBON CANYON RANCH RAVEN PROJECT PROGRESS REPORT, DECEMBER 2001 John P. Kelly Cypress Grove Research Center Audubon Canyon Ranch, Marshall, CA 94940 kellyjp@svn.net Jennifer E. Roth Point Reyes Bird Observatory 4990 Shoreline Highway, Stinson Beach, CA 94970 ACR Tech. Rpt. 98-9-2 December 2001, AUDUBON CANYON RANCH 4900 Shoreline Highway Stinson Beach, CA 94970

2001 ACR RAVEN PROJECT PROGRESS REPORT 1

2001 ACR RAVEN PROJECT PROGRESS REPORT 2 Audubon Canyon Ranch Raven Project Progress Report, December 2001 INTRODUCTION The modification of natural landscapes by humans has led to changes in habitat quality and quantity that adversely affect many avian species, while other species benefit from such changes. For example, corvid populations are increasing throughout much of the United States, apparently because of their ability to exploit a variety of humanproduced food sources. In addition, many corvids are opportunistic predators on other avian species and have the potential to adversely affect their populations. This effect is especially pronounced in depleted populations such as the Snowy Plover in Point Reyes National Seashore, which declined 32% from 1986-2000 largely because of nest predation by ravens (Ruhlen and Abbott 2000, Point Reyes National Seashore report). Recent increases in the number of Common Ravens (Corvus corax) in the San Francisco Bay area have coincided with more frequent observations of ravens in heron and egret nesting colonies. In 1998, Great Egrets at the Bolinas Lagoon Preserve of Audubon Canyon Ranch (ACR) suffered severe nest predation by resident ravens. Most nests were lost when nestlings reached 3.5-5 weeks of age and only 30 young were successfully fledged, compared with an expected production of 100-150 young based on annual monitoring since 1967. In 1999 and 2000, the ravens at ACR nested in or near the heronry trees, but failed to raise young ravens. During those years, raven predation

2001 ACR RAVEN PROJECT PROGRESS REPORT 3 in the colony was low and the egrets and herons produced a normal number of chicks. In 2001, ravens nested in upper Pike County Gulch, about 1 km from the heronry, where they successfully raised two chicks. They continued to prey on egret nests, but nest predation was not heavy. We have also observed annual raven predation at Marin Islands National Wildlife Refuge, near San Rafael. This site is one of the most important heronries in the San Francisco Bay area and supports several hundred nests of Great Egrets, Snowy Egrets, Black-crowned Night-Herons, and Great Blue Herons. Ravens were first observed at this heronry in 1993. Since then, ravens have nested annually on East Marin Island and preyed on heron and egret nests on West Marin Island. These and other observations around the Bay Area suggest an urgency to understand the potential threat of raven predation to waterbird colonies. Raven predation at heronries could become a more widespread problem as ravens become more prevalent. If ravens become a regional problem, knowledge about local raven behavior could become valuable in protecting heron and egret populations. However, it is not known whether raven predation threatens populations of herons or egrets or has other significant regional effects. In response to concerns about increasing raven predation on other species, ACR and the Point Reyes Bird Observatory are conducting a multifaceted study of ravens in the northern San Francisco Bay area. The project involves evaluation of home range size, habitat use, and movement patterns using radio telemetry; investigation of raven behaviors at selected heron and egret nesting colonies; regional analysis of population status and distribution of ravens and crows; and experimental attempts to achieve "conditioned taste aversion," a potential management technique that uses illnessproducing prey to alter nest predatory behavior.

2001 ACR RAVEN PROJECT PROGRESS REPORT 4 ACCOMPLISHMENTS How does home size range vary among individuals and locations? Home Range analysis is based on radio telemetry results, including ground and aerial surveys to track movement patterns and habitat use in Marin and Sonoma counties. Estimates of home range are being modeled for each individual raven under study, so that individual differences in home range, movement patterns, and habitat use can be evaluated. Table 1 indicates the level of radio telemetry effort to date. Table 1. Radio telemetry effort on Common Ravens in west Marin County. Category 1999 2000 2001 Total # captured 16 14 2 32 # fitted with radios 13 12 2 27 # individuals tracked 10 17 16 22 # dead 2 2 3 9 # failed transmitters 1 1 2 # left area 2 2 # nests monitored 6 11 14 31 # successful nests 2 7 9 18

2001 ACR RAVEN PROJECT PROGRESS REPORT 5 The field data gathered to date are adequate to complete the home range analysis, although additional telemetry work is being conducted to further understand patterns of movement (see below). We are using kernal home range and minimum convex polygon estimation techniques to estimate the size of 16 individual home ranges for which we have adequate data. To examine intraseasonal changes in raven movements and habitat use, we plan to evaluate relationships between nesting stage (e.g., incubation, nestlings, post-fledging) and home range size. Mean annual survivorship of raven nests was 54% ± 10 (SE; Table 1). Habitat selection by ravens is being evaluated in relation to the availability of each habitat type within each bird s home range and across the study area. These analyses involve 60-120 radio telemetry locations for each bird, estimates of individual home range boundaries, and use of vegetation layers available in existing geographic databases. Preliminary analysis suggests that many of the home ranges were within 2-6 km 2, with considerable variation among individual pairs. In 1999, we produced a report of preliminary findings and management recommendations to the Point Reyes National Seashore (Roth, J. E., Kelly J. P. Sydeman, W. J, Parker, M. W., and Allen S. A. 1999. Ecosystem Management of Common Ravens on the Point Reyes National Seashore. 20 December 1999, 25 pp.) What is the status and distribution of regional raven and crow populations? An analysis of the distribution and abundances of Common Raven and American Crow in the San Francisco Bay area was recently completed. We have produced an ACR Technical Report on the study (Kelly and Etienne 1999, Abundance and distribution

2001 ACR RAVEN PROJECT PROGRESS REPORT 6 of Common Raven and American Crow in the San Francisco Bay area: results of a 1999 road survey. ACR Tech. Rpt. 98-9-1.), and a scientific paper has been submitted for peer-review publication (Kelly, Etienne, and Roth, Abundance and distribution of Common Raven and American Crow in the San Francisco Bay area, California. Western Birds in review). The study was based on a road survey conducted in 1999 by 40 of ACR s volunteer field observers, and on other available data from the North American Breeding Bird Survey, Christmas Bird Counts, and breeding bird atlas projects in eight San Francisco Bay counties. Although both ravens and crows occur throughout our region, the results indicate that ravens concentrated along the outer coast and occur in relatively low numbers in some interior areas (but are abundant in others). In contrast, crows were relatively scarce along the outer coast, increasing significantly in abundance in interior and Bay shore locations. Both species occurred in greater densities along urban and suburban survey routes than along rural routes, but dramatic exceptions were evident in some areas. For example, crows occurred in unusually high numbers in the rural farmland northeast of Suisun Bay, and ravens occurred in relatively low numbers in urbanized east Marin County. The highest densities of ravens occurred on the Point Reyes Peninsula. In an effort to help reduce raven predation on Snowy Plover reproduction, the Point Reyes National Seashore recently cited our road surveys (Kelly and Etienne 1999) and our report to the National Seashore (Roth et al. 1999) in a proposal to modify ranching practices associated with high concentrations of ravens (Biological assessment on renewal of livestock grazing permits, Point Reyes National Seashore, July 12, 2001). Regional abundances of ravens and crows have increased strongly over the last 20 years. This is certainly no surprise for species that benefit from agriculture, road kills,

2001 ACR RAVEN PROJECT PROGRESS REPORT 7 and garbage in human-altered landscapes. However, analysis of available data on ravens and crows for the San Francisco Bay area revealed that localized trends within the region vary substantially and even include areas with declining numbers of ravens. These differences in local trends and abundances of ravens and crows suggest that explanations for increasing corvid populations may not be so simple. Overall increases in the numbers of ravens and crows in the San Francisco Bay area may involve a range of local differences in habitat suitability, or alternatively, localized opportunities for further population growth. What behavioral patterns characterize raven nest predation at heronries? Qualified volunteer field observers have monitored more than 60 heronries in the northern San Francisco Bay area for reproductive success and raven activity since 1999. Raven behaviors were observed intensively at selected heron and egret colonies, with field effort concentrated at sites with resident ravens such as the heronries at West Marin Island, near San Rafael, and ACR s Picher Canyon at Bolinas Lagoon. Four hundred eighty-three 2-hr observations of raven behaviors at heron and egret colony sites were obtained (Table 2). These observations were gathered during early, midday, and late periods to allow investigation of diurnal differences in the use of various behaviors. During each 2-hr observation period, detailed accounts of behaviors were recorded whenever a raven landed in the colony, interacted with another bird, or flew over the colony. In addition, observers recorded the instantaneous state of raven activity at 5-min intervals. This allowed us to quantify the extent of overall raven activity with regard to the number of ravens, their locations in or near the colony site, their use of

2001 ACR RAVEN PROJECT PROGRESS REPORT 8 various perching substrates, and their positions relative to available habitat structure in the nesting colony. The data are being used to estimate the proportion of time ravens were present, patterns of raven behavior associated with nest predation and disturbance to nesting birds, and the importance of habitat features to ravens in the vicinity of the colony. Table 2. Preliminary summary of behavioral observations of ravens at selected heron and egret colonies. No. of 2-hr Interactions / hr Year observations Landings / hr Fly-overs / hr Dominant Subordinate 1999 155 0.93 1.01 0.10 0.11 2000 180 1.84 1.10 0.17 0.08 2001 148 2.83 1.61 0.16 0.10 In the analysis of raven behaviors, we are examining differences associated with time of day, intraseasonal timing, raven nesting stage, and heron and egret nesting stage. We are also examining relationships among different raven behaviors to determine overall patterns of activity near heronries. Preliminary results from the first season of field work indicate that ravens occupied the nesting colony at Marin Islands 25.7% of the time (SE=5.8), with 1.38 individuals (SE=0.06) present, on average, during each occurrence; ravens occupied the egret colony at ACR's Picher Canyon 6.5% of the time (SE=1.2), with 1.23 individuals present, on average, during each occurrence. At

2001 ACR RAVEN PROJECT PROGRESS REPORT 9 these and other sites, the pair of resident nesting ravens occupied the vicinity of each nesting colony almost continuously. Ravens patrolled the colonies regularly, and often scanned nests in the colony from live branches above the tree canopy. We observed relatively few actual predation events, on both eggs and nestlings. The frequency of observed colony landings and fly-overs by ravens seems to have increased over the course of this study (Table 2). A more detailed analysis of behavioral data is continuing. Regional results indicate that most heronries in the region are not significantly disturbed by ravens, but that some ravens specialize on heron and egret colonies. We are currently analyzing the overall effects of raven predation on nest survivorship at selected sites. At ACR s Picher Canyon, ravens destroyed at least 28 Great Egret nests during the egret nestling period in 2001, representing a loss of at least 56 nestlings. Egg predation at ACR has been relatively rare. At West Marin Island, we quantified egg predation rates using an index based on weekly searches for depredated eggs on East Marin Island where ravens nested. Preliminary results suggest a dramatic increase in Table 3. Number of depredated eggs found on East Marin Island, 1999-2001. Year No. weekly search-days Heron and egret eggs Other eggs* Total eggs 1999 8 45 10 55 2000 3 16 16 32 2001 6 140 31 171 *Other eggs include Western Gull and Mallard eggs.

2001 ACR RAVEN PROJECT PROGRESS REPORT 10 egg predation at West Marin Island in 2001 (Table 3). In addition, observations of ravens suggested an increase in the frequency of raven behaviors inferred or suspected to be directly associated with mortality of heron or egret nests or nestlings at West Marin Island and Picher Canyon (Table 4). These behaviors include ravens flying from the colony with eggs, young, or unidentified food, and ravens observed in or adjacent to a failed nest. Survivorship of focal Great Egret nests at West Marin Island (69%) was also Table 4. Frequency of raven behaviors inferred or suspected to be directly associated with mortality of heron or egret nests or nestlings at West Marin Island and Picher Canyon, 1999-2001. Number of nest predatory behaviors / 100 hr Colony site 1999 2000 2001 West Marin Island 6.52 13.50 17.78 Picher Canyon 2.00 4.63 7.77 Total 3.06 6.90 10.81 lower in 2001 (69%, n = 54 focal nests) than in 2000 (81%, n = 59) or 1999 (80%, n = 45), which was consistent with an apparent increase in nest predation by ravens. Anecdotal evidence of predation by ravens, based on carcasses found near raven roosts on East Marin Island, further indicated predation of adult Snowy Egrets by ravens: at least seven in 2001 and at least four in 2000 (none found in 1999). We strongly emphasize that the results presented here are preliminary. We are currently conducting

2001 ACR RAVEN PROJECT PROGRESS REPORT 11 a more rigorous analysis of trends in raven predation and survivorship of heron and egret nests at the West Marin Island and Picher Canyon heronries. What management alternatives are likely to be most successful in controlling raven predation at heronries? We attempted to establish "conditioned taste aversion" (CTA) in ravens at ACR s Picher Canyon by providing chemically treated prey that can produce severe illness in ravens, and possibly, alter their predatory behavior. This part of the project required long periods of observation to identify the source of all nest failures in the egret colony, and close monitoring of raven behaviors before, during, and after the experiment. We collaborated on the experiment with Dr. Lowell Nicolaus of Northern Illinois University, who has authored several scientific papers on CTA (e.g., Nicolaus 1987. Am. Midl. Nat 117: 405-419; Nicolaus and Lee 1999. Ecol. Appl. 9(3): 1039-1049). In 1999 and 2000, the resident pair of ravens did not approach the egret carcasses or other food items provided to them during the CTA tests possibly because of reduced food requirements related to their own nest failure. In 2001, we were able to capture both of the resident ravens at Picher Canyon. We radio-tagged them and provided them with Great Egret meat treated with fenthion, which they both consumed ad libitum in captivity. The birds were then released, but no subsequent behavioral changes were detected. Tests of CTA cannot be replicated at a single heronry because only one pair of ravens occupies each site. Replication would require substantially more effort, to treat of additional ravens and detect changes in

2001 ACR RAVEN PROJECT PROGRESS REPORT 12 rates of nest predation at other heronries. Therefore, we were not able to develop a fully controlled experiment, and in interpreting the apparent absence of a change in raven behavior, we could not rule out the possibility of an error in CTA application or dosage. Our current knowledge of the CTA process includes the following information. CTA has been established in free-ranging Common Ravens as well as in free-ranging American Crows, resulting in their avoidance of preying on eggs (Nicolaus et al. 1983. Science, 220:212-214; Nicolaus 1987. Am. Midl. Nat. 117:405-419). In addition, CTA-treated ravens have excluded non-territorial ravens from nesting areas of Sand Hill Cranes and Least Terns (Nicolaus 1987; Avery et al. 1995. Colonial Waterbirds 18:131-138). To avoid confusion in the taste-aversion process, only the appropriate prey (taste) should be used in treatment. In addition, CTA is most likely to be successful if it is achieved with the first meal of the day (soon after dawn). Newly captured ravens readily consume food provided to them ad libitum after only 15-30 minutes, providing humans are not present (birds at ACR were observed through a remote video monitor). Captive feeding should not adversely affect the CTA process because the primary stimulus is taste, and is transmitted directly through a medullar pathway. Other conditions by which ravens might distinguish treated food from actual prey, such as visual cues, location, or reasoning, are secondary to conditioning process. After

2001 ACR RAVEN PROJECT PROGRESS REPORT 13 treatment, birds should learn quickly to associate actual predation attempts with the taste that stimulates the aversive response. Ravens are difficult to catch, often requiring two or more weeks of daily baiting and several additional days to capture them. Capturing individuals a second time is possible, but may be more difficult. Treating free-ranging ravens to establish CTA is difficult. Most free-ranging ravens will not readily consume treated bait provided for them, especially if they suspect that human observers are involved. However, observers are needed to confirm that the food was consumed by a raven rather than by another species, and to prevent non-target species from taking the food. Ravens may require two weeks or considerably longer to become accustomed to a potential treatment area. Evidence of human activity should not be detectable at or near the treatment area, and the food must be provided just before dawn. Food provided to the ravens must be fresh and provide the appropriate taste. Therefore, if achieving CTA requires an extended effort, it could be difficult to obtain enough food (e.g., egret meat). To administer an adequate dose, the treated food must be consumed completely by the target individual, not shared with its mate or offspring or taken by other species. Other options for controlling nest predation by ravens suggest little promise. The use of visual, auditory, or chemical repellants to discourage raven predation is unlikely to

2001 ACR RAVEN PROJECT PROGRESS REPORT 14 succeed beyond an initial response period, and suggests potential disturbance to nesting herons and egrets. There is no consistent empirical support for the use of raven carcasses or models as effigies to deter raven activity. Removing ravens from the site by shooting, trapping, or poisoning, would present considerable difficulties. Because ravens are extremely wary, they are very difficult to shoot and shooting both members of a nesting pair may not be possible. In addition, the safe use of firearms with regard to human activity in the vicinity of the ACR heronry substantially limits firing angles and positions. Shotguns require close range, which is rarely available with ravens. Rifles allow a greater range but carry additional restrictions for safe use and require greater accuracy. Trapping may be the most feasible method of removal, but would require substantial effort at ACR because the birds have already been captured once and retrapping both individuals of a pair would be difficult. Ravens cannot be translocated because of potential problems with disease transmission and exporting pest species to other areas. Removal by poisoning would require exclusion of herons, Turkey Vultures, Scrub Jays, Owls, and other native species from poisoned bait. Any removal of ravens would probably require an ongoing control program as other ravens move in to fill vacancies. If a new pair of resident ravens did not immediately fill the vacancy, vagrant non-territorial ravens might prey on heron and egret nests at equal or greater levels. Nest predation at ACR s heronry on Bolinas Lagoon has been reduced in years when raven nesting attempts have failed. Therefore, disrupting successful nesting by ravens might help in controlling nest predation. Possible methods of disrupting raven nesting behavior include the removal of nests, eggs, or young, or treating eggs so they will not hatch. Addling (by shaking), oiling, or puncturing eggs to prevent hatching has the potential advantage of delaying the ravens detection of nest failure, and thus delaying or preventing subsequent renesting. However, their likely response to such

2001 ACR RAVEN PROJECT PROGRESS REPORT 15 treatments is unknown and it may be impossible to approach a raven nest without being detected. Disruption of successful nesting would require annual searches for nest locations and follow-up searches for renesting locations within seasons. Ravens may initiate two or more renesting attempts in one season. Because the resident ravens at ACR are radio-tagged, the potentially difficult task of locating nesting sites should be relatively simple. A qualified tree climber would be required. If ravens abandon the area after nest disturbance, a new nesting pair might establish a territory, or in the absence of a new pair, vagrant groups of ravens might occupy the undefended area. How do ravens structure their daily movements and use of habitats? Radio-telemetry observations used to estimate home range were gathered by recording three independent locations of each marked bird each week during each nesting season. Therefore, such data are of limited use in evaluating diurnal movement patterns characteristic of each individual. To measure movements of individual birds, we are now tracking the locations of individual ravens at 5-min intervals for periods of 3-6 hrs. These data will provide at least three kinds of information. First, they will allow us to look for consistent daily routines in their use of habitats within their home range. Second, we will be able to better quantify the extent of daily movement and areas of intensive use. Finally, these data will allow us to estimate the temporal scale within which ravens structure their movements. Preliminary analysis suggests that raven movements may be predictable within fairly short periods of time (approximately 90

2001 ACR RAVEN PROJECT PROGRESS REPORT 16 minutes), after which they have an equal chance of moving anywhere within their home range. Further work is needed to address these questions. PLANS FOR 2002 FIELD SEASON In 2002, we will concentrate efforts on home range analysis, behaviors at heronries, diurnal movements, and management options, including CTA. We are analyzing home ranges for 16 nesting ravens that were tracked in 2000 and 2001, and identifying patterns that may account for differences in home range size and habitat use among these individuals. We are also analyzing patterns of raven activity at heronries, based on nearly 500 2-hr observations gathered from 1999 to 2001. We will evaluate these patterns using log-linear models that account for correlated effects among variables. We are devoting considerable field time during winter 2001-2002 and spring 2002 to following radio-tagged ravens for periods of 3-6 hours. The results will be used to quantify the diurnal movement patterns of individual ravens. If possible, we will try again to establish CTA In 2002. However, because freeranging ravens at ACR were not willing to consume treated food during 1999 and 2000, we may not be able to treat them for CTA again unless we can capture them a second time. We are also investigating the possibility of establishing CTA in resident ravens at the large heronry at Marin Islands National Wildlife Refuge (Bryan Winton, personal communication). We plan to collaborate further with Lowell Nicolaus, on controlled experiments to measure CTA in captive ravens in eastern Oregon, to establish a more accurate prescription for future field efforts. Finally, given difficulties in finding feasible

2001 ACR RAVEN PROJECT PROGRESS REPORT 17 methods controlling raven predation at heronries, we are continuing to investigate other options for reducing raven predation at ACR s Picher Canyon heronry, including disruption of raven nesting behavior.