Document Details
 |
Title |
Testing relationships between habitat quality indices, forest configuration and Marbled Murrelet local population size | ||||||
Author |
Forest Investment Account (FIA) | ||||||
Date |
2009 | ||||||
Abstract |
Marbled Murrelets are seabirds that nest in the canopies of large seral conifers up to 50 km from the coast. The species is listed as Threatened south of Alaska, primarily due to loss of nesting habitat. Maintaining sufficient habitat throughout coastal British Columbia is a priority for long-term sustainability, under the assumption that habitat is a surrogate for the population size of birds potentially supported within an area (Burger 2001). In five regions of BC, radar-derived indices of murrelet abundance during the breeding season correlate remarkably strongly, and apparently linearly, with the amount of old growth forest per watershed (Burger 2001), and with areas of three forest-cover based definitions of murrelet suitable habitat (Burger et al. 2004a). The strategic goal in managing terrestrial habitat for Marbled Murrelets is achieving target areas of ?suitable habitat? in each of six conservation regions (Canadian Marbled Murrelet Recovery Team, 2003). However, the shape of the relationship between local population size and habitat area and quality are issues with major impacts on forest management, as well as questions about the nesting productivity value of habitat with respect to fragmentation and landscape context (Waterhouse et al. 2002, 2004,2007, in review(a,b); Burger et al. 2004b; Zharikov et al. 2006, 2007a,b; Burger and Page 2007, Malt 2007, Malt and Lank 2007; Bahn and Lank, in revision). Two six-rank habitat quality classifications based on interpreting stand structure from air photos and helicopter surveys are currently used to spatially define nesting potential. Air photo classifications are available for Haida Qwaii, and will be available for the Central Coast (2008) and North Coast (2009). Helicopter surveys will be available for North and West Vancouver Island, parts of the Southern Mainland, and Mid-coast in 2008. Both types of mapping are thought to improve the reliability of polygon boundaries and refine estimates of habitat availability. This information is directly incorporated into forest management, but a persistent issue is how to value areas of different rank when establishing wildlife reserves. There is conflicting evidence about the effects of forest edges on murrelet nest success. If fragmentation effects vary regionally, by elevation, or change over time, because of differences in predator densities, these effects can be incorporated into harvest and conservation planning. Previous FSP-sponsored research tested the extent to which habitat quality indices predicted real murrelet nest sites, and supported experiments on edge effects. However, these studies could not quantify the actual relationships between murrelet density and habitat quality and configuration, which would directly link landscape and population management. Our proposed analyses test the value of four measures of habitat quality as predictors of local population size, as indexed by radar counts of commuting birds taken during the breeding season. Predictor variables are areas of: spatially explicit ?habitat suitability?, as assessed by a province-wide strategic map of dichotomous suitability classification similar to that used by Burger et al. (2004b), (2) ranks of habitat derived from air photo interpretation (3)ranks based on low-level helicopter surveys, and (4) ?edge vs core? suitable and/or ranked habitat, and the landscape matrix within which habitat occurs, which depend on habitat shape, size, and configuration. We will make use of radar data not available for earlier analyses, use updated classifications and GIS databases of potential habitat, and assess the utility of novel analytical approaches to model these relationships in at least seven areas of British Columbia, distributed among at least four Marbled Murrelet management regions. For watersheds where radar and habitat rankings exist, we will examine whether quality class distributions provide better predictive models of radar counts than ... | ||||||
Report Number |
FIA2009MR032 | ||||||
|
|||||||
