12/14/96

 

AN EVALUATION OF THE IMPACT OF DIFFERENT LAND MANAGEMENT
SCENARIOS ON THE POPULATION OF THE MARBLED MURRELET
IN HUMBOLDT COUNTY, CALIFORNIA

 

C. John Ralph and Sherri L. Miller
U.S.D.A. Forest Service, Redwood Sciences Laboratory
Arcata, California 95521

 



INTRODUCTION

 

An innovative large-scale exchange of lands has been proposed for certain coastal Redwood forests in Humboldt County, California. The region involved includes portions of the drainages of the Eel, Van Duzen, and Elk Rivers, which we refer to below as the 'bioregion'. This report represents the first step in an evaluation of the effect of such a transfer on Marbled Murrelet habitat and populations. Our goal here is to determine the effect of the existing agreement on the murrelet.

Two complementary approaches will be carried out in tandem to assess the overall strategy of the exchange. The first analysis, reported herein, will focus on the conservation and management of habitat, and will determine the likely population size under different preservation and harvest scenarios. The second analysis, the Population Viability Analysis (Akçakaya et al., in prep) will concern population viability under these different conditions.

It should be emphasized that the present report is the initial step in a process that will entail much further analyses. The results reported here must be viewed as preliminary, and should be used only as a broad indication of the consequences of the proposed exchange. They are primarily a road map to the ongoing analyses. We have made many simplifying assumptions in this initial study, and have largely restricted ourselves to use of existing data sets compiled by Pacific Lumber Company. As the analyses proceeds during 1997, we will refine and delimit these assumptions, and include data from a variety of other sources. We anticipate receiving feedback from reviewers during these subsequent steps, and incorporating a wider range of conditions, in an iterative process.

 

Our basic procedure is as follows:

 

  1. The extent of existing Marbled Murrelet habitat of differing qualities in the bioregion will be determined for different ownerships, including those in the proposed exchange.
  2. The frequency of Marbled Murrelet detections on surveys in the various forest habitats and densities in adjacent marine waters will both be used to estimate the current use of different forest stands. In effect, the population will be allocated to different forest stands on the basis of observed behaviors.
  3. This resulting product will then allow us to estimate the effect of preservation and loss of habitat under different management scenarios, and hence the potential effects on the Murrelet population. It will also allow us to assess the potential augmentation of suitable habitat through active management or succession.

We emphasize that this basic approach incorporates several major assumptions. We anticipate, for instance, relaxing the assumption that all stand conditions are effectively equal in suitability. For instance, the assumption that an isolated, fragmented stand, far from the ocean, is equivalent to an unfragmented stand of similar size, close to other habitat and the ocean. In addition to unweighted allocation, we will incorporate landscape and other criteria to weight the allocation of murrelets to different stands. Similarly, we will evaluate the probable future conditions of different stands under plausible scenarios of forest succession, blow-down, fire events, landslides, and other stochastic events. A third initial assumption is that birds in harvested areas are lost to the population. This may be untrue, as birds may be able to relocate elsewhere.

The final product of these analyses will then be an evaluation of the likely consequences of the land exchange, under a range of different conditions. Our goal is not to produce a single best estimate of such efforts, but to develop a comprehensive approach, so that decision makers can evaluate alternative options. These conditions will include both optimistic and pessimistic scenarios and assumptions.

 

 


ESTIMATE OF SUITABLE HABITAT

 

To estimate the amount and types of suitable murrelet nesting habitat included in the Pacific Lumber Timber Property (the "Property") outlined in the agreement, criteria for potential habitat within the region were selected and habitat stands were defined. The criteria for potential habitat on the Property were selected from characteristics of stands surveyed since 1988 and found to have murrelets present or exhibiting occupied behaviors and included:

 

  1. unharvested old-growth redwood/Douglas-fir, or
  2. residual (partially harvested) redwood/Douglas-fir stands, with at least 25% canopy closure provided by the residual old-growth trees and/or a minimum 40-year-old second growth component.

All stands to be surveyed were inspected to verify habitat typing. The Property includes 24 stands of potential murrelet habitat. These stands are identified in Figure 1 and listed in Tables 1 and 2 and total approximately 6,298 acres. Additional potential habitat may exist in small scattered fragments and these will be identified in Stage II. Potential habitat on State Park lands in the region was identified by interpretation of aerial photographs using similar characteristics, but has undergone limited inspection (Fig. 1). Suitable habitat on the park lands totaled 20,310 acres. In addition, all federal and other private timberlands in the region are shown in Figure 1 (figure available from authors).

We are continuing with a descriptive analysis at the stand and landscape scales. Habitat characteristics will be summarized for each stand from measurements obtained from vegetation plots at murrelet stations and from estimates from aerial photography interpretation. Landscape features at the patch and landscape levels will be examined using Fragstats (1994), software developed for analysis of spatial patterns in the GIS environment. Variables examined may include: core area, contagion, shape indices, juxtaposition and interspersion.

We will use a variety of statistical methods to identify habitat variables with possible effects on murrelet distribution or densities. The relationship of murrelet activity and habitat characteristics will include univariate and multivariate techniques, such as, principal component and logistic regression analyses. Potential habitat designation may be reevaluated following the habitat relationships analyses.

 

 


ESTIMATION OF HABITAT AND POPULATION
IN PROPOSED CONSERVATION AREAS

 

We estimated the proportion of the Murrelet population that currently occurs on lands that will be set aside by the following methods:

 

 

  1. Conservation areas were set as the Headwaters and Elkhead Springs stands, as well as the Humboldt Redwoods State Park. Management areas were considered as all other areas on Pacific Lumber Company lands with occupied murrelet detections.
  2. The acreages of suitable habitat in the State Park were then estimated under two acreage alternatives, one with abundant habitat availability of 20,310 acres, or moderate habitat availability of 4,500 acres.
  3. The proportion of the bird's population in the habitat areas was then estimated from the land acreages, either unweighted, or weighted by bird abundance. Bird abundance was estimated by either the average number of detections at stations in the stand, or by the average number of occupied behaviors seen at stations. Weightings for State Park lands were based on 1994-95 surveys at Bull Creek only. Weightings for Pacific Lumber lands are based on average detection levels throughout the stands, as summarized in data provided by Pacific Lumber Company.

In this preliminary analysis, the percentage of the murrelet population that will be protected is estimated at between 71.1% and 98.0% (see summary Tables 3a and 3b). Highest protection levels were predictably found under the assumptions that a proportionally large amount of the area of suitable habitat occurs on State Park lands, and that these areas have relatively high occupancy levels. Lowest protection levels were found for the scenario where acreages alone were used (unweighted by murrelet behavior), and moderate levels of habitat were assumed for the Park.

We separately analyzed the proportion of murrelets, currently on Pacific Lumber lands, that would be protected under the proposed plan. This calculation is not affected by habitat levels on the park, but can be considered under weightings for murrelet behavior. The percentage of murrelets currently on Pacific Lumber lands, whose habitat would be preserved under the plan, is estimated between 50.4% and 77.3%. Highest protection was found for the scenario where acreages were weighted by detection levels. Lowest protection was found under the scenario where acreages were unweighted.

We emphasize that these results are preliminary. The original data will be reviewed, compiled, and will be reanalyzed in Stage II of this project. Additional data that are available for other sites in the Park and elsewhere will be incorporated in this reanalysis. We anticipate that this could well result in reduced occupancy levels in the State Park, with resultant reduced protection levels there. Results for protection on Headwaters and Elkhead Springs, relative to other Pacific Lumber lands, will likely be unaffected, but the proportion of birds on this land may change.

Stage II of this project will also include more refined estimates of the amount and distribution of habitat present on the State Park. These data may also modify our results. We will continue to present a range of scenarios, to cover all likely outcomes.

 

 

 


INLAND MURRELET DETECTIONS

 

For the process of weighting the suitability of habitat across the Property, the results of inland surveys for murrelets from 1994, 1995, and 1996 were summarized by Pacific Lumber Company for each stand and provided to us. Surveys conducted in 1994 and 1995 in the Humboldt Redwoods State Park also were summarized. Surveys were conducted during the nesting seasons, according to the Pacific Seabird Group's Inland Survey Protocol (Ralph et al. 1994) recommendations. Stands with detections were considered "occupied" (Ralph et al. 1994) if within canopy behaviors were observed. The average number of detections and observations of occupied behavior were calculated by stand, and various weightings used, to allocate a proportion of the total birds to each stand (Tables 1 and 2).

To refine the estimated proportions of the population in each stand during Stage II, we will apply results from the habitat analyses to adjust for effects of habitat parameters on likely bird densities in each stand. A system of weighting any unsurveyed areas of the stands will incorporate possible differences in habitat quality within the stands. Detection levels will then be extrapolated to unsurveyed areas according to habitat quality, the total detections for the region summed, and the proportion of detections for each stand recalculated. The proportion of detections will then be applied to the estimated total birds in the region, by the methods in Ralph et al. (1995), to estimate the number of birds using individual stands.

The accuracy of our estimates will be assessed by the rigor of meeting various assumptions. The assumptions will include: (1) accuracy of habitat mapping; (2) a relationship between habitat area and patterns of detection levels; (3) the likelihood of detections being directly correlated to numbers of birds in the stand; and (4) the ability of observers to detect birds, given varied station placement within each stand. We will identify and test, when possible, the assumptions we have made and the effects of any error in the assumptions.

 

 


OFFSHORE POPULATION

 

It is not possible to directly count the number of birds associated with an individual forest stand. However, by using the methods outlined above, we can estimate the number of birds associated with stands from relative stand detection levels and then relate these to the estimated offshore population for the bioregion.

To define the population that may be nesting in the bioregion, we considered two factors. Observations of radio-tagged murrelets indicate that nesting birds may fly up to 50-70 km from nest sites (Jodice pers. comm.) to forage in marine waters. When extrapolated from the bioregion these distances encompass the coastal section of California from approximately Redwood Creek to Cape Vizcaino. Density estimates in this section range from 0.72 to 8.81 birds per km2 (Ralph and Miller 1995). The highest densities are in the northern part of the section, and the lowest in the south. Within this section we looked for significant changes in density from north to south. By post hoc examination, we selected the coastal section from Trinidad to Shelter Cove as representing the birds potentially nesting within the bioregion. Although some of the birds in the most northerly and perhaps the most southerly parts of the section are likely to be nesting outside of the bioregion, we feel this would be offset by some birds from the bioregion foraging to the north or south of this section and therefore not being included. We have estimated (Ralph and Miller 1995) a total of 1,694 murrelets in the coastal section from Trinidad to Shelter Cove. The range of standard error of this population estimate was very low, 14.9 to 72.8.

In Stage II, we will continue to refine this population estimate, using additional data. Further, we will examine any trends in population in the area over the past eight years.

Finally, we will examine the methods used to estimate the number of potential nesting birds in the bioregion area.

 


EFFECT OF PRESENT LANDSCAPE AND ENVIRONMENTAL FEATURES

 

Landscape level features will be incorporated in the second stage of this analysis. The present (Stage I) study, assumes that all habitat stands are equally affected by variable landscape characteristics. The final product of Stage II will then be a weighting of stand suitability to take account of landscape level features. This will also incorporate possible future changes in such features. Habitat stands will be characterized on various attributes, detailed below. The data will be modeled using both a geographic scale and a metric-based analysis.

Geographic information systems data will be used, and analyzed with the statistical package FRAGSTATS, as well as similar packages.

 

Geographic location.--Several geometric factors are likely to have an influence upon the effectiveness of a given stand in being suitable for nesting of murrelets. Data for these analyses will be both from Pacific Lumber Company and U.S. Forest Service GIS databases.

 

  1. Juxtaposition.--the proximity of each stand to other, nearby stands is a critical factor in the stand's potential to act as murrelet nesting habitat. The effect of such weightings will be to increase the relative importance of stands close to other habitat, and to decrease the importance of isolated stands.
  2. Shape and core area.--the average distance of a stand from the edge is a important metric of the potential exposure of nesting birds to predators. Compact stands with high core to edge ratios can be weighted higher than dissected stands with relatively large amounts of edge.

Aggregation of habitat stands.--The effect on population density of several stands in a relatively small area (e.g. the Headwaters complex), vs. scattered, smaller stands, is likely to be important. Those stands that are aggregated can be weighted higher than scattered stands.

Distance from ocean.--A simple variable, yet one that is likely to be vital, is the distance of each stand to the ocean. This is a metric that reflects the energy demands for birds to reach nesting areas. Distance to the ocean can be used as a weighting so that stands farther from the ocean are ranked lower.

Stochastic events.--While difficult to model, it is likely that certain events or circumstances will affect the rate of succession, as well as the aggregation of suitable habitat. These include potential for blow-down, fire, and corvid predation. Predation and blow-down are both, for instance, more likely in stands with much edge. Stands which are more prone to disturbance (such as wind and human-caused fires) could be weighted lower than less prone stands.

Behavioral response to harvesting.--Two types of response can be anticipated. First, an immediate response to disturbance or harvesting, at least at a suitable distance, does not appear to be critical (Long and Ralph, in prep). Thus restrictions on land use, based on this criterion, may not be very important. At this point, we do not propose any weighting of stands in response to nearby harvest activities. On the other hand, long-term displacement is very likely to have an effect. Murrelets are likely to be displaced from harvested stands and may pack into nearby remaining suitable habitat, resulting in an increased density in occupied stands. This can result in potentially negative effects, such as the use of substandard nest-sites and increased predation. It can also have potential positive effects, such as increased sociality or synchrony, or the elimination of low quality sites. In the initial analysis here, we have assumed that all birds displaced from harvested lands are lost. This assumption will be relaxed in Stage II, so that displaced birds may be transferred to nearby stands, albeit with the potential of increased predation risks, enhanced sociality, and other factors.

 

 

 


LONG TERM HABITAT CHANGES

 

The current analysis considers only short-term changes to murrelet habitat of less than 25 years. It does not incorporate any changes that can be anticipated in the longer term of 50 years and beyond. The final stage of the analysis will include such changes as follows:

 

Successional changes
In redwood forests, murrelets are mostly associated with old-growth, although some individuals occur in stands with residual trees. It is expected that such residual stands may well become more attractive to murrelets as successional changes occur and the surrounding forest develops. It may also be that some areas of second-growth develop sufficient structure to attract nesting murrelets. The development of new habitat is expected to occur only in conservation areas: the State Park, and the Headwaters complex. We will estimate the effect of such successional changes on the development of habitat in the two areas. Succession is likely to increase the value of State Park and other conservation lands.

Succession also is affected by the forest matrix in which the conservation areas are embedded. Even if the matrix does not become habitat itself, it can promote murrelet breeding success by reducing predator efficiency, eliminating edge-effects, wind-throw, and other effects. Such effects will be incorporated in Stage II of the analysis.

 

Silvicultural promotion of habitat
The current analysis of Stage I assumes that the Headwaters and Elkhead Springs stands will become conservation areas, and that all other habitat on Pacific Lumber lands will be harvested. It also assumes that no promotion of habitat succession, on Pacific Lumber lands or elsewhere, will occur. Silvicultural prescriptions that promote the development of old-growth characteristics are thought to be possible. For instance, thinning and other practices can encourage both tree growth and the development of lateral limb structure, as is being proposed for Redwood National and State Parks (1996). We will evaluate the potential of such techniques to accelerated production of murrelet habitat in the second stage of this analysis.

 

Risk of catastrophic events in conservation centers
There exists a risk that the Headwaters complex, or the State Park conservation area, will be lost or reduced due to catastrophic events. We will in part consider these risks in our analysis, they will also be fully discussed in the Population Viability Analysis.

 

 


CONCLUSIONS

 

All conclusions of this preliminary analysis are subject to modification during the Stage II of the investigation, and are primarily based upon Pacific Lumber Company data.

 

 

 

Table 1. Marbled Murrelet Habitat on State Park (BullCreek survey) and Pacific Lumber lands. Shown are the size of stands, the average number of Murrelet detections/visit (obs/visit), the average number of occupied detections / visit (ocobs/visit), and the percentage of Murrelet habitat in the bioregion in each stand under three scenarios (acreage only, acreage weighted by detections, and acreage weighted by occupied detections). Also shown is the percentage of Marbled Murrelet habitat on Pacific Lumber lands represented by each stand, under the same three scenarios.

These data are proprietary data; any use without permission is a breach of professional ethics. Contact C. John Ralph, U.S.D.A. Forest Service, Redwood Sciences Laboratory, Arcata, California 95521

 

% Birds in bioregion, weighted by: % Birds on Pacific Lumber Lands, weighted by:
Stand Acres Obs / visit OcObs / visit None Obs OcObs None Obs OcObs
Bull Creek 4500 9.56 2.70 41.67 43.72 73.63 NA NA NA
Headwaters 2755 14.81 1.06 25.51 41.47 17.70 43.74 73.69 67.11
Elkhead 422 4.69 0.58 3.91 2.01 1.48 6.70 3.57 5.62
Allen Creek 391 3.52 0.29 3.62 1.40 0.69 6.21 2.49 2.61
Copper M 363 1.88 0.18 3.36 0.69 0.40 5.76 1.23 1.50
Shaw Gift 294 6.46 0.5 2.72 1.93 0.89 4.67 3.43 3.38
Bell Law 286 3.44 0.51 2.65 1.00 0.88 4.54 1.78 3.35
Owl Cr A 245 1.13 0.29 2.27 0.28 0.43 3.89 0.50 1.63
Bemis A 231 2.98 0.51 2.14 0.70 0.71 3.67 1.24 2.71
Bemis B 231 2.47 0.44 2.14 0.58 0.61 3.67 1.03 2.33
Nanning 229 16.19 0.36 2.12 3.77 0.50 3.64 6.70 1.89
Jordan 159 4.18 0.08 1.47 0.68 0.08 2.52 1.20 0.29
N Fork Elk R 159 0.59 0.02 1.47 0.10 0.02 2.52 0.17 0.07
Road 3a 154 3.17 0.12 1.43 0.50 0.17 2.45 0.88 0.64
Booths R 109 6.81 0.11 1.01 0.75 0.42 1.73 1.34 1.60
Right 9 76 0.57 0.03 0.70 0.04 0.01 1.21 0.08 0.05
Rd 11 Ba 48 1.92 0.36 0.44 0.09 0.10 0.76 0.17 0.40
Rd 11 Bb 48 0.63 0.08 0.44 0.03 0.02 0.76 0.05 0.09
Lower Rd 2 24 2.11 0.06 0.22 0.05 0.01 0.38 0.09 0.03
Below Rd 7 21 0.24 0.06 0.19 0.01 0.03 0.33 0.01 0.11
Rd 12 La 18 5.33 1.00 0.17 0.10 0.58 0.29 0.17 2.20
Rd 12 Lb 18 3.52 0.07 0.17 0.06 0.38 0.29 0.11 1.46
Road 3b 10 2.89 0.11 0.09 0.03 0.18 0.16 0.05 0.66
Chadd Cr 8 1.27 0.09 0.07 0.01 0.06 0.13 0.02 0.23

Table 2. As Table 1, but assuming 20310 acres of habitat in the State Park.

Marbled Murrelet Habitat on State Park (BullCreek survey) and Pacific Lumber lands. Shown are the size of stands, the average number of Murrelet detections/visit (obs/visit), the average number of occupied detections / visit (ocobs/visit), and the percentage of Murrelet habitat in the bioregion in each stand under three scenarios (acreage only, acreage weighted by detections. and acreage weighted by occupied detections). Also shown is the percentage of Marbled Murrelet habitat on Pacific Lumber lands represented by each stand, under the same three scenarios.

These data are proprietary data; any use without permission is a breach of professional ethics. Contact C. John Ralph, U.S.D.A. Forest Service, Redwood Sciences Laboratory, Arcata, California 95521

 

% Birds in bioregion, weighted by: % Birds on Pacific Lumber Lands, weighted by:
Stand Acres Obs / visit OcObs / visit None Obs OcObs None Obs OcObs
Bull Creek 20310 9.56 2.70 76.33 77.81 92.65 NA NA NA
Headwaters 2755 14.81 1.06 10.35 16.35 4.94 43.74 73.69 67.11
Elkhead 422 4.69 0.58 1.59 0.79 0.41 6.70 3.57 5.62
Allen Creek 391 3.52 0.29 1.47 0.55 0.19 6.21 2.49 2.61
Copper M 363 1.88 0.18 1.36 0.27 0.11 5.76 1.23 1.50
Shaw Gift 294 6.46 0.5 1.10 0.76 0.25 4.67 3.43 3.38
Bell Law 286 3.44 0.51 1.07 0.39 0.25 4.54 1.78 3.35
Owl Cr A 245 1.13 0.29 0.92 0.11 0.12 3.89 0.50 1.63
Bemis A 231 2.98 0.51 0.87 0.28 0.20 3.67 1.24 2.71
Bemis B 231 2.47 0.44 0.87 0.23 0.17 3.67 1.03 2.33
Nanning 229 16.19 0.36 0.86 1.49 0.14 3.64 6.70 1.89
Jordan 159 4.18 0.08 0.60 0.27 0.02 2.52 1.20 0.29
N Fork Elk R 159 0.59 0.02 0.60 0.04 0.01 2.52 0.17 0.07
Road 3a 154 3.17 0.12 0.58 0.20 0.05 2.45 0.88 0.64
Booths R 109 6.81 0.11 0.41 0.30 0.11 1.73 1.34 1.60
Right 9 76 0.57 0.03 0.29 0.02 0 1.21 0.08 0.05
Rd 11 Ba 48 1.92 0.36 0.18 0.04 0.03 0.76 0.17 0.40
Rd 11 Bb 48 0.63 0.08 0.18 0.01 0.01 0.76 0.05 0.09
Lower Rd 2 24 2.11 0.06 0.09 0.02 0 0.38 0.09 0.03
Below Rd 7 21 0.24 0.06 0.08 0 0.01 0.33 0.01 0.11
Rd 12 La 18 5.33 1.00 0.17 0.04 0.58 0.29 0.17 2.20
Rd 12 Lb 18 3.52 0.07 0.17 0.03 0.38 0.29 0.11 1.46
Road 3b 10 2.89 0.11 0.09 0.01 0.18 0.16 0.05 0.66
Chadd Cr 8 1.27 0.09 0.07 0 0.06 0.13 0.02 0.23

 

Table 3a. The percentage of the Marbled Murrelet population estimated to be contained in proposed protected areas (State Park, Headwaters, Elkhead Springs) under various assumptions.

Habitat acreage in the park is evaluated under two alternative scenarios (4,500 and 20,310 acres of suitable habitat). Murrelet populations are then assumed to vary simply with acreages, or are weighted by detection levels or by occupied detection levels. Weightings for park surveys are based on 1994-5 surveys in Bull Creek only.

These data are proprietary data; any use without permission is a breach of professional ethics. Contact C. John Ralph, U.S.D.A. Forest Service, Redwood Sciences Laboratory, Arcata, California 95521

Weightings on Acreages derived from:
Habitat in Park (acres) Unweighted Detections Occupied Detections
20,310 88.27 94.94 97.98
4,500 71.09 87.20 92.81




Table 3b. The percentage of Marbled Murrelets on Pacific Lumber lands that are estimated to occur on proposed protected lands (Headwaters and Elkhead Springs), under three scenarios. Murrelet populations are assumed to vary simply with acreages, or are weighted by detection levels or by occupied detection levels.

These data are proprietary data; any use without permission is a breach of professional ethics. Contact C. John Ralph, U.S.D.A. Forest Service, Redwood Sciences Laboratory, Arcata, California 95521

Weightings on Acreages derived from:
Unweighted Detections Occupied Detections
50.44 77.26 72.78


LITERATURE CITED

 

 

Long, L.L. and C.J. Ralph. in prep. Effect of human disturbance on nesting Marbled Murrelets, Alcids, and other seabirds.

Ralph, C.J.; S.K. Nelson, M.M. Shaughnessy, S.L. Miller, T.E. Hamer, and Pacific Seabird Group, Marbled Murrelet Technical Committee. 1994. Methods for surveying Marbled Murrelets in forests. Technical paper #1, revision. Available from: Oregon Cooperative Wildlife Research Unit, Oregon State University, Corvallis, OR; 48 p.

Ralph, C.J. and S.L. Miller. 1995. Offshore population estimates of Marbled Murrelets in California. pp. 353-360 in Ecology and Conservation of the Marbled Murrelet. Gen. Tech. Rep. PSW-152. Albany, CA: Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture.

Ralph, C.J., S.L. Miller and L. Folliard. 1995. An estimate of population size of an offshore and forest population of Marbled Murrelets in northwestern California. Final report to Arcata Redwood Company, Orick, California.

Redwood National and State Parks. 1996. 'Second-growth Forest Recovery Plan and Environmental Assessment'. Unpublished draft, September. Crescent City, California.