Northern saw-whet owl (Aegolius acadicus brooksi) COSEWIC update and status report: chapter 7

• Previous page
• Table of contents
• Next page

Population Sizes and Trends

Search effort

To date, three sets of surveys have been conducted on the Haida Gwaii subspecies of Northern Saw-whet Owl. First, Gill and Cannings (1997) did a systematic survey in 1996, where 238 survey stations across a variety of habitat types on Graham and northern Moresby Island were sampled twice during the spring breeding season, leading to the detection of 61 owls. Second, CIH undertook two years of targeted searches for Saw-whet Owls (2002/2003) on the southern half of Graham Island in maturing and old forested habitats. In total, 24 and 26 occupied sites were found in 2002 and 2003 respectively (Holschuh 2004). Third, CIH undertook a systematic survey of Gwaii Haanas National Park Reserve in 2004, where survey stations were accessed by boat and located in protected inlets, passages and bays, and targeted adjacent terrestrial habitats, to ultimately determine range and density. Of the 59 stations surveyed (many of which were surveyed twice), Saw-whet Owls were found at 26 sites (Holschuh in prep.). The detection success was highest in the Gwaii Haanas survey. These three sets of surveys were done during spring when all males are highly vocal.

For the purposes of this document, population estimates were derived from measures of density and extrapolated to the amount of habitat available currently and ten years ago. Two population densities were calculated: (1) the density of active owl sites on Graham Island in a landscape altered by forestry activities using the 2002/2003 database, and (2) the density of active owl sites in a large tract of undisturbed habitat, in Gwaii Haanas. As these owls appear to be associated with mature and old forests, estimating the density of owls in a largely undisturbed landscape should provide a preliminary baseline estimate with which to compare densities in a changing landscape. Densities were calculated based on average distance to the closest neighbour. Raptors often have a uniform distribution across the landscape. The approximate spacing between territories can be taken as a surrogate to home range size. If the distance between the territory cores (areas the males are most consistently defending) is divided by two, a circular home range area could be calculated. By then dividing the calculated home range size into the amount of available habitat, the population size could be estimated. As the population estimate comes from the average spacing between individuals, the error in the population estimate was calculated from the standard error of the mean distance between individuals.

There are numerous limitations to this method. First, because the specific habitat needs of these owls are not fully understood, calculating home ranges, and then population size based on the amount of available habitat may lead to overestimating the population size. Although the density measures may take factors such as fragmentation into account, other habitat features may also be limiting factors. Also, not all available habitat will necessarily be used by owls. Secondly, after estimating the number of sites across the landscape, this number is being multiplied by two. This is making an assumption of equal sex ratio, which has not been verified. Another limitation is that the population is being estimated based purely on habitat availability, and cannot take into account other possible factors, such as fecundity. Density estimates were extrapolated from data collected in different years, when annual variation in population size is not known. Ultimately, changes in the availability of habitat are being taken as surrogates to population change, when there are likely additional factors that are not yet understood (i.e. reproductive success changes over time due to the changing landscape). This could be leading to an overestimate in number of sites and individuals. Overall, the population sizes and trends have not been measured directly. Despite these weaknesses, this method is applying measured densities from survey data to changing habitat availability, thereby providing a reasonable estimate of abundance.

Abundance

Using the above estimates, the density of owls within Gwaii Haanas in contiguous old forest is significantly greater than the density in fragmented regenerating forest. Surveys in contiguous forests in Gwaii Haanas produced a mean of 2.18 + 0.18 km between occupied sites, and therefore an approximate mean home range size of 3.52 + 1.3 km². Dividing this figure into the 2,561 km² calculated for total unfragmented suitable forest (see Habitat) produces a total of 725 + 200 territories in unfragmented landscapes. Conversely, in fragmented habitat, the average inter-neighbour spacing was 4.31 + 0.71 km, producing a mean home range size estimate of 14.52 + 11.1 km². Again, using the 2,927 km² of available habitat in fragmented landscapes, this translates into a total of 201 + 90 territories in this habitat type. The combined total of 926 + 290 territories is a best estimate for the number of possible breeding territories on Haida Gwaii, representing a possible adult population of 1,852 + 580 adult owls.

Fluctuations and trends

There is very little information for either subspecies on population fluctuations and trends. Although there is some suggestion that populations cycle with food abundance in A. a. acadicus (Cannings 1993 and references therein), the more opportunistic and generalist nature of foraging in A. a. brooksi (Sealy 1999) seems to suggest some imperviousness to the fluctuations of single prey items. Nonetheless, repeated surveys of the same site between years showed that areas of concentrated activity in one year may have much lower activity in the next year (Holschuh, unpub. data).

When the rates of decline in unfragmented habitat are taken into account, and the spacing of owls (density) is extrapolated to this changing landscape, a 9% decline in the population size of Saw-whet Owls over the past 10 years can be estimated. This is based on a harvest rate of old forests of about 3,000 ha per year (Figure 5), which translates into a loss of habitat for about 85 pairs of owls over 10 years. The loss over the last 15 years, the presumed upper estimate for three generations, would be about 13%. Further, if the population is projected on the original base of available unfragmented habitat, it suggests there may once have been approximately 1,550 + 420 territories, suggesting an approximately 40% decline in population since the start of large-scale forest harvest operations (which began about 1950, Gowgaia Institute data, A. Cober, pers. comm.). Overall, as the amount of unfragmented mature and old forest continues to be lost on Haida Gwaii, the number of owls can also be expected to continue declining. Further, the effects of other external factors, including the introduction of exotic species, are not well understood and are not factored into these calculations. Factors such as these are likely having an effect on the population also, suggesting that the above estimates are conservative. The collective effects of the limiting factors and threats described below will likely cause further decline of Haida Gwaii Saw-whet Owls.

Forest harvest is projected to continue at about 3,000 ha per year on the Queen Charlotte Islands, almost all occurring in old-growth forests (Alvin Cober, pers. comm.). This would translate into a decline of approximately 6% in the Saw-whet Owl population (57 pairs lost in a population of 926 pairs) over the next 10 years, or 9 % over 15 years (upper limit of estimate for three generations).

Rescue effect

As the brooksi subspecies of the Northern Saw-whet Owl is endemic to the Haida Gwaii archipelago, there is no rescue effect.