Home range and resource selection by animals constrained by linear habitat features: an example of Blakiston's fish owl

Authors


Summary

  1. Typically in resource selection studies, the spatial extent of a home range is defined first and then the available resources within that perimeter are estimated. However, the home ranges (or habitats) of some animals are constrained by linear environmental features (e.g. rivers, shorelines). Traditional home range estimators often overestimate home range extent for such species, which can lead to spurious estimation of resource availability and selection.

  2. We used a synoptic model of space use to explicitly account for resource selection of a species constrained by linear features in its environment to compare with traditional home range estimators. We used the endangered Blakiston's fish owl Bubo blakistoni in the Russian Far East as our example.

  3. Mean annual home range size (± standard error) was more than three times larger when using kernel methods (30·3 ± 15·1 km2) than when using the synoptic model (9·4 ± 2·0 km2, = 7).

  4. Fish owls showed strong selection for areas within valleys, closer to waterways, closer to patches of permanently open water and with greater channel complexity than available sites.

  5. Synthesis and applications. The synoptic model solves a long-standing problem in home range and resource selection studies because it provides an objective way to estimate the space use of a species whose habitat is constrained by linear features in its environment. Improvements in the accuracy of such estimations can lead to identification of important resources across landscapes, the development of more rigorous site-specific or landscape-scale management plans, and to scientifically defensible conservation or threat mitigation measures.

Introduction

Estimation of resource selection by a species is an important tool for conservation planning (Manly et al. 2002). Traditionally, to estimate patterns of resource selection within an animal's home range, the perimeter of the home range is defined using an appropriate estimator (e.g. kernel density; Worton 1987); then, habitat or other resources within the perimeter are quantified and compared with their relative availability (e.g. Singleton et al. 2010; Vanak & Gompper 2010).

Home range models that estimate a utilization distribution (e.g. kernel density estimators or bivariate normal distributions) are useful because they quantify a species' probability of use of a given location (Jennrich & Turner 1969; Horne & Garton 2006). Kernel density estimators are particularly attractive because home range estimates using these methods often stabilize with ≤50 telemetry locations, which makes them popular when data are limited (Kernohan, Gitzen & Millspaugh 2001). However, some species use habitat in ways that cannot be satisfactorily described by standard utilization distribution methods (e.g. species whose ranges are constrained by linear features in a landscape, such as canyons, rivers or coastlines). In such cases, most estimators overestimate home range size because they include large areas of unused habitats (Samuel & Fuller 1996; Horne, Garton & Rachlow 2008). This problem is so pronounced for telemetry studies of fish that some authors omit home range estimates entirely, thereby ignoring potentially valuable information on space use (Knight et al. 2009). Consequently, attempts have been made to address the limitations of traditional estimators, including manipulation of the parameters used to define kernel density estimators (Sauer, Ben-David & Bowyer 1999), use of cluster analysis (Kenward et al. 2001; Knight et al. 2009) or k nearest-neighbour convex hull (Getz & Wilmers 2004), inclusion of animal movement data (Rhodes et al. 2005) and a synoptic model of space use that can use covariates to help define the boundaries of a home range (Horne, Garton & Rachlow 2008).

We assessed the suitability of the synoptic model (Horne, Garton & Rachlow 2008) to describe space use of Blakiston's fish owl Bubo blakistoni (hereafter ‘fish owl’). As presented by Horne, Garton & Rachlow (2008), the synoptic model (so named because it estimates several metrics simultaneously) is a model of space use that incorporates both home range and resource selection processes. Integration of home range and resource selection is a recent trend in analytical approaches (e.g. Johnson et al. 2008; Forester, Im & Rathouz 2009; Aarts, Fieberg & Matthiopoulos 2012). However, prior to our study, the synoptic model had not yet been adopted and applied to a species-specific study. This may be due partly to the lack of software packages for implementing newer approaches, but equally likely is that the generality or flexibility of the synoptic model has not yet been recognized. Given that the fish owl presumably uses river and stream habitats almost exclusively (Surmach 1998; Yamamoto 1999), we specifically examined fish owl resource selection to evaluate the general efficacy of this model to resolve a long-standing problem with estimating home range size and resource selection by species occupying linear environments.

Materials and methods

Study Area

Our 20 213-km2 study area was located on the eastern slope of the Sikhote-Alin Mountains, in Primorye Province, Russia (see Fig. S1 in Supporting Information). The study area covered ~12% of Primorye, extending from the Avvakumovka River drainage (near the village Olga; 43°43′23″N, 135°15′20″E) north to the Maksimovka River drainage (north of the village Amgu; 45°50′27″N, 137°40′40″E). Elevations ranged from 0 to 1733 m. Temperatures were warmest in August (mean temperature = 22 °C) and coldest in January (mean temperature = −20 °C; Newell & Wilson 1996).

Sample Population

We captured fish owls from eight territories in three geographical regions (Olga, Ternei and Amgu; see Fig. S1 in Supporting Information) from 2007 to 2010 primarily using the methods of Slaght, Avdeyuk & Surmach (2009). Briefly, we placed live fish or frogs in an enclosure set within a stream to lure owls to a trap set near the enclosure. We marked adult, territorial fish owls in 2008–2010 with GPS data loggers (40 and 90 g models, Sirtrack Tracking Solutions, Havelock North, New Zealand), which were attached with 1·10 cm Teflon-coated ribbon as a harness (Bally Ribbon Mills, Pennsylvania, USA) following methods of Kenward (2001). Data loggers required recapture of the birds to recover location data. Therefore, any animal that died or dispersed resulted in loss of data. We programmed 90 g data loggers to record one location every 11 h and 40 g units every 23 h to conserve battery power. All GPS data loggers weighed ≤3% of both sexes (mean male mass was 3·10 ± 0·03 kg [mean ± SE, = 7 owls], and mean female mass was 3·25 ± 0·12 kg [= 5]), which was within recommended limits (Kenward 2001). We estimated GPS location quality using GPS fix success rate, proportion of 2-D vs. 3-D fixes and location error (Cain et al. 2005). We eliminated GPS locations with 2-D fixes and those with horizontal dilution of precision values >5 because they may be unreliable (Dussault et al. 2001). We partitioned our data into four phenological seasons (winter = 1 December–31 March, spring = 1 April–14 June, summer = 15 June–14 September and autumn = 15 September–30 November; see Slaght 2011).

Data Analysis and Model Development

We estimated home range size and within-home range resource selection (equivalent to Johnson's (1980) ‘third-order selection’) using a synoptic model of space use (Horne, Garton & Rachlow 2008; example code provided in Appendix S1 in Supporting Information), which ameliorated some limitations of standard home range estimators by allowing estimation of both home range and resource selection within the same modelling process. The structure of the synoptic model was based on a weighed distribution that is used to predict a species’ probability of use of a given area (Lele & Keim 2006; Horne, Garton & Rachlow 2008; Johnson et al. 2008). Each individual's space use (i.e. home range) was modelled as follows:

display math(eqn 1)

where fu(x) is the probability density of being at spatial location x, fo(x) is the null distribution of space use which models the probability of use in the absence of habitat selection and w(x) is a selection function that transforms fo(x) to fu(x)by selectively weighing different areas based on habitat attributes at location x. For our analysis, we defined fo(x) = N (θ), a bivariate normal distribution with five parameters, symbolized by θ, describing the means and variances in the x and y dimensions and the covariance. We used a bivariate normal distribution as our null model, as this distribution links use with a central place (Horne, Garton & Rachlow 2008). This is an appropriate (i.e. biologically meaningful) null model for fish owls because they are nonmigratory and centre their territories on a nest tree used for multiple seasons (Slaght 2011). We defined the selection function as

display math(eqn 2)

where H (x) is a vector of covariate values describing the habitat at location x and β is a vector of parameters (i.e. selection coefficients) to be estimated. To accommodate temporally changing habitat covariates (e.g. distance to open water), we generalized eqn (eqn 2) to depend on the values of the covariates at time t. Thus, the time-specific resource selection function was

display math(eqn 3)

If all areas were equally available across the landscape, eqn (eqn 3) would be analogous to a ‘resource selection function’ (i.e. a function that is proportional to the probability of a resource unit being used given knowledge of its habitat characteristics; Manly et al. 2002). However, the synoptic model allows the values of the availability distribution fo (x) to vary across the landscape. Thus, while eqn (eqn 3) is similar in spirit to a resource selection function (i.e. it multiplicatively transforms the availability distribution into the use distribution), its interpretation as well as the associated values of the selection coefficients must include consideration of the availability distribution (see explanation below).

The value of the estimated selection coefficients can be interpreted via probability ratios. In general, we want to understand how the probability of occurrence would change at a given location x with covariate values H(a) compared with an alternative location with covariate values H(b). The probability ratio quantifying how much more or less likely an animal is expected to be at a location with covariate values H(a) vs. covariate values H(b) is

display math(eqn 4)

where K is the integral in the denominator of eqn (eqn 1). Importantly, eqn (eqn 4) assumes location x is equally available for subsequent selection in both hypothetical areas a and b (i.e. the role played by fo (x) in both the numerator and denominator). For example, suppose elevation is one of the covariates in our model with values that range from 1000 to 3000 in our study area. For estimation purposes, we standardize the covariate to range from 0 to 1, and we subsequently estimate the selection coefficient to be inline image . To interpret this value, we are interested in how much more likely the animal is to occur at an elevation of 2200 vs. 2000 (i.e. standardized values of 0·6 and 0·5, respectively). From eqn (eqn 4), we can calculate this value as

display math

Thus, given equal availability (i.e. accessibility) within the home range, the animal is twice as likely to be at an elevation of 2200 vs. 2000. Note that this value will be the same for any 200-unit increase in elevation (e.g. 200 vs. 0, 850 vs. 650, etc.). Also note that because eqn (eqn 4) is based on the ratio of probabilities, this interpretation corrects the assertion by Horne, Garton & Rachlow (2008) that eqn (eqn 4) is an ‘odds ratio’.

We used maximum likelihood to estimate the parameters governing the null model of home range (θ) and the selection coefficients(β). We approximated the integral in eqn (eqn 1) by placing an ‘availability grid’ (a grid of points) spaced 50 m apart over each fish owl region (mean size was an 818 295-point grid across a 2111 km2 area; = 3). We chose this grid size as a trade-off between spatial resolution and processing speed.

We developed habitat covariates based either on our knowledge or the literature on fish owls (Slaght & Surmach 2008; Slaght 2011). We used a covariate (VALLEY) to assess important landform features within the home range (Takenaka 1998; Slaght 2011). VALLEY was a continuous variable estimated using a topographic position index (TPI) in Land Facet Corridor Designer (a module in ESRI ArcGIS; v. 9.3, Redlands, CA, USA), where negative values indicate valleys locations, values near 0 indicate low slopes, and positive values indicate high slopes or ridge tops (Jenness, Brost & Beier 2012). In our study landscape, TPI values distinguished valleys (TPI < −0·62) from slopes (−0·62 < TPI < 0·90) and ridges (TPI >0·90). We also generated a covariate depicting neighbouring owl territories (NEIGH), which we created with a minimum convex polygon to bound home ranges of neighbouring owls; otherwise, the synoptic model would include all nearby suitable habitat, which is not realistic for this territorial species.

We created seven continuous covariates: distance to water (D_WAT), distance to patches of permanently open water; that is, river stretches that remain unfrozen in winter (D_WWAT), amount of deciduous forest (DEC_213), amount of riparian old-growth forest (RO_213) and open space (i.e. young forest, field, road, talus slope; OPEN_213) within a 213-m buffer. We also measured the amount of open water in winter (WWAT_213). We included a covariate for channel complexity (i.e. number of river channels; CHAN_213) and one interactive term (RO*DWAT), which evaluated the interaction between riparian old-growth forest and distance to water covariates. Fish owls appeared to be closely associated with water (Slaght & Surmach 2008); therefore, we chose a 213-m buffer to sample resources because this was the mean distance (plus two standard deviations) of all fish owl locations to water.

We used remotely sensed data to create a cover type map of the study area. We identified five cover types: coniferous forest, deciduous forest, riparian old-growth forest, water (river, lake) and open space (bare surfaces, settlements, agriculture, other open areas). We used SPOT five imagery (Satellite Pour l'Observation de la Terre) because high spatial resolution (10 m) facilitated mapping open water (particularly in winter) and some cover types. We used 30 m Landsat TM data (Sioux Falls, SD; http://glovis.usgs.gov/) from September 2007 because we could not reliably distinguish old-growth forest using SPOT. We enhanced Landsat TM images for analysis using a variety of techniques and assessed map accuracy following Congalton & Green (1999) with some modification (see Slaght 2011). Overall classification accuracy was 85%, with a Kappa statistic (a measure of agreement between our classified map and reference data; Titus, Mosher & Williams 1984) of 0·82 (see Slaght 2011). We conducted all remote sensing analyses using ESRI ArcGIS and ERDAS Imagine (v. 9.3; Norcross, GA).

To assess within-home range resource selection by fish owls, we developed 14 a priori additive and interactive models using the above nine covariates (see Slaght 2011). Prior to analysis, we tested covariates for colinearity and removed highly correlated covariates ( 0·7). In such cases, we retained the covariate that made the most sense, given our knowledge of fish owl biology and ecology. Half of the models (50%, = 7) were variants of what we called the ‘literature’ model, which included the covariates VALLEY, D_WWAT and RO_213. An additional four models were variants of what we called the ‘multi-channel’ model, which had similar components to the ‘literature’ model but also included D_WAT and CHAN_213.

We selected the best synoptic model from the candidate set based on Akaike's information criterion corrected for small sample sizes (AICC; Burnham & Anderson 2002; Horne, Garton & Rachlow 2008). We drew inference from models based on AICC weights (Burnham & Anderson 2002) and strength of parameter estimates (85% confidence intervals [CI]; Arnold 2010). If individual fish owl data sets had sufficient sample size (i.e. if the best model was anything but the null model [bivariate normal distribution]), we included them in seasonal analysis. We excluded data for home range size analysis from an unpaired owl from autumn because its movement patterns indicated that it did not have a stable home range. However, we retained its location data for parameter estimation of resource selection.

We averaged parameter estimates from all top models across individuals to generate seasonal parameter estimates and averaged seasonal estimates to derive an annual estimate (similar to ‘two-stage’ methods; Fieberg et al. 2010). For annual estimates, we weighed parameter estimates for each season by its relative proportion of the year (i.e. spring was 75 days long, so spring selection coefficients were weighed by 0·21).

We modelled six additional home range estimators including the exponential power distribution, bivariate normal, a two-mode bivariate circular normal mix, a two-mode bivariate normal mix, and fixed kernel and adaptive kernels. We determined the best of these using the likelihood cross-validation criterion (CVC; Horne & Garton 2006), in program Animal Space Use (v. 1.3; Horne & Garton 2007). Computation time, involving numerical maximum likelihood, prevented us from calculating CVC for the synoptic model. However, to provide a general comparison of the performance of the synoptic model vs. the alternatives, we compared the AICC score of the best synoptic model with the CVC score of the best alternative estimator (Stone 1977; Horne & Garton 2006). We executed the synoptic model using program R (v. 2.11, R Development Core Team, www.r-project.org) using code written by D. Johnson and J.S.H.

Results

We captured 16 fish owls (seven adult males, six adult females and three juveniles) of the 17 we detected during four field seasons (January–May), totalling 20 months of capture effort. We marked eight adult resident owls from six territories in four river drainages with GPS data loggers. One of these owls was not detected the year following capture, so no data were retrieved. We recorded 1892 locations from seven fish owls on five territories and recorded 28–192 locations per fish owl per season (mean = 90·2 ± 10·1; see Table S1 in Supporting Information). Mean GPS fix success was 0·6 ± 0·1 (range: 0·2–1·0) for all birds across all seasons. By season, fix success was highest in winter (mean = 0·8 ± 0·1; range: 0·7–0·9), followed by spring (mean = 0·7 ± 0·1, range: 0·3–1·0), autumn (mean = 0·5 ± 0·1, range: 0·2–0·8) and summer (mean = 0·5 ± 0·1, range: 0·3–0·7).

Home Range and Resource Selection

Mean annual home range size using kernel smoothing was 30·3 ± 15·1 km2 (= 7; see Table 1 for seasonal estimates). In contrast, mean annual home range size using the synoptic model was 9·4 ± 2·0 km2. AICC values from the synoptic model were lower than CVC values for kernel estimates in 71% of analyses (= 15; see Table S2 in Supporting Information). The mean distance between the farthest locations within a home range for all owls was 12·4 ± 32·4 km (Table 1).

Table 1. Seasonal and annual home range estimates for individual Blakiston's fish owls in Primorye, Russia, from data collected 2008–2010. Home ranges varied by season and were much smaller using the synoptic model as compared to kernel density methods for all seasons except winter. Annual home range size derived from the synoptic model was also much smaller than its kernel density counterpart
BirdWinterSpringSummerAutumnAnnual
SynaKernbLincSynKernLinSynKernLinSynKernLinSynKernLin
  1. a

    Syn = estimates derived from the synoptic model (km2).

  2. b

    Kern = estimates derived from either fixed or adaptive kernel density methods (km2).

  3. c

    Lin = estimates of maximum linear distance between points (km).

  4. d

    Annual estimate based on two or three seasons.

Sha-Mi F2·12·02·76·14·69·46·68·813·08·022·912·05·79·69·3
Sha-Mi M12·19·714·56·76·09·67·831·012·714·116·618·010·215·813·7
Kudya F4·05·98·119·515·111·614·817·014·59·530·115·012·017·012·3
Kudya Md1·72·73·310·97·27·66·35·05·5
Saiyon Md1·012·57·12·411·17·71·711·87·4
Sereb Md3·19·56·715·346·215·618·747·619·112·434·413·8
Faata Md24·9151·720·710·185·229·417·5118·525·1
Average5·05·17·210·329·510·49·533·215·512·629·316·09·430·312·4
SE2·41·82·73·320·41·82·011·93·02·26·01·42·015·12·4

No single model emerged as clearly superior as a result of the model selection process. Five different models were identified as the best fit across 21 analyses, with three of these models accounting for 81% of all top models (= 17; see Table S3 in Supporting Information). Two models were the top model in 29% of analyses (= 6); the first was a variant of the ‘multi-channel’ model (including the covariates VALLEY, D_WAT, RO_213 and CHAN_213), and the second contained covariates VALLEY, RO_213, DEC_213 and OPEN_213. The next most common model contained the covariates VALLEY, RO_213, DEC_213 and D_WWAT, which accounted for 24% of all top models (= 5).

Certain covariates appeared regularly. For example, the covariate D_WAT was found in 76% of top models (= 16), despite being in only 50% of all models, and the covariate CHAN_213 was present in 43% of top models (= 9), despite being in only 36% of all models.

Parameter Estimation

Of the eight covariates in top models that were related to fish owl occurrence, only D_WAT (annual β = −18·0 ± 4·3) and CHAN_213 (annual β = 1·4 ± 0·3) had parameter estimates that did not overlap 0 across all seasons and annually. In addition, VALLEY (annual β = −7·0 ± 1·7) had an annual estimate that did not overlap 0, but had a seasonal estimate (winter) with less statistical support. Lastly, D_WWAT (annual β = −19·7 ± 8·0) had an annual estimate that did not overlap 0, but was not present in any top model in autumn, so the annual estimate is based only on three seasons (see Table S4 in Supporting Information).

Discussion

Space use by fish owls is constrained by habitat that is arranged linearly in their environment (i.e. riparian forest), and the synoptic model provides an objective method to estimate their home range and resource selection accurately. In our analysis, this model outperformed competing kernel density estimators both visually (see Fig. 1) and quantitatively (see Table S2 in Supporting Information). It therefore has application for developing recommendations for fish owl conservation, and it has the potential to be used for estimating space and resource use by other wildlife taxa whose space use is constrained by linear features within landscapes.

Figure 1.

Example of cumulative probability output of the synoptic model overlaid on adaptive kernel density home range estimator for a female Blakiston's fish owl with = 121 GPS locations at the Sha-Mi territory, in Primorye, Russia, summer 2009.

Synoptic Model vs. Traditional Home Range Estimators

The synoptic model was initially developed to combine two related processes (i.e. home range behaviour and resource selection) into a single modelling framework (Horne, Garton & Rachlow 2008). We demonstrated here that this method also solves the long-standing problem of home range estimation for species whose habitat is constrained by linear environmental features (e.g. canyons, rivers). Studies of resource selection by such species have been confounded by the shortcomings of traditional home range estimators because estimates derived from traditional estimators usually included large areas as potentially used habitat (Blundell, Maier & Debevec 2001; Knight et al. 2009). Hence, home range size was usually overestimated. Several solutions for this problem have been presented in the literature (e.g. Sauer, Ben-David & Bowyer 1999; Blundell, Maier & Debevec 2001; Kenward et al. 2001; Knight et al. 2009), but none appear satisfactory (Slaght 2011). The synoptic model of space use apparently offers an elegant solution. Whereas strict utilization distribution-based methods such as kernel density estimators approximate home range using only the spatial distribution of animal locations, the synoptic model uses animal locations in conjunction with the environmental covariates associated with those locations (Horne, Garton & Rachlow 2008). Animal home ranges will often be shaped by topography and other environmental features (Powell & Mitchell 1998), so their inclusion into the spatial definition of home range is intuitive. Given that appropriate environmental covariates will be selected to model space use, and that a biologically meaningful null model of space use is selected, this method will have a wide range of applications across taxa.

Most home range estimates derived using the synoptic model had utilization distributions that appeared more biologically realistic than kernel density estimates (i.e. they did not include habitats that owls were unlikely to use). This assessment was supported quantitatively by a comparison of AICC and CVC values (see Table S2 in Supporting Information). The synoptic model's home range included predicted use in areas outside the distribution of GPS locations. It did this because, unlike kernel density estimators that define home range irrespective of environmental influences, the synoptic model predicted the presence of likely habitat near fish owl locations.

Parameter Estimates and Interpretation

The Blakiston's fish owl is a species whose home range is defined by linear features (waterways) in its environment; a statement supported by very strong annual selection coefficients for distance to water (D_WAT; β = −18·0 ± 4·3) and landform (VALLEY; β = −7·0 ± 1·7; Table S4 in Supporting Information). In conjunction with channel structure (CHAN_213; β = 1·4 ± 0·3) and distance to patches of permanently open water (D_WWAT; β = −19·7 ± 8·0), these covariates helped the synoptic model outperform kernel density estimators by guiding probability of use estimation (Fig. 1).

Probability ratios showed that, for example, a fish owl was 6·3 times more likely to be found 53 m from water (the mean distance to water based on GPS locations) than 753 m (the maximum distance) and was also 2·7 times more likely to be found near a waterway with 6 channels (the maximum value based on GPS locations) than a waterway with only 1 channel. Finally, fish owls were 1·9 times more likely to be found within the valley landform (where TPI = −0·90; the mean value of GPS locations) than on a slope (TPI = 0·10) and 5·1 times more likely to be found within the valley landform (where TPI = −1·63; the minimum value of GPS locations) than on a ridge (TPI = 1·0). The minimum and maximum covariate values used to generate probability ratios are found in Table S5 in Supporting Information.

Seasonal Variation in Resource Selection

Although our differences in home range size across seasons suggested variation in resource use, quantifying these differences was complicated by small seasonal sample sizes and different rates of GPS fix success. We believe that lower fix success in autumn and summer may have been related to increased foliage cover during these months. Low GPS success rates in autumn and summer precluded conclusive statements about home range shifts during those seasons; however, generally, high fix success in winter and spring, when fish owls did not exhibit the extreme movements of summer and autumn, supported the inference that fish owls showed differences between season in home range and resource use. Seasonal differences in space use and GPS fix rate may also be due to distribution and movement of prey, which may have increased use of smaller waterways by fish owls. Such tributaries are used by spawning salmonids in summer and autumn and have denser canopy cover than larger waterways, which would not allow as clear a view of the sky for GPS fixes.

Conservation Implications

Fish Owls

Yamamoto (1999) recognized that traditional home range estimators were not appropriate for fish owls, but no reasonable alternative was available. Results using the synoptic model confirmed his observation because we found that previous estimates of fish owl home range area were imprecise (e.g. Hayashi 1997; Takenaka 1998; Yamamoto 1999; Andreev 2009). The range of annual and seasonal movements we described revealed strong association with riparian areas within valleys that had complex river systems – a quantitative confirmation of the importance of these resources as suggested by natural historians (e.g. Pukinskii 1973; Mikhailov & Shibnev 1998). Although the generality that fish owls use such areas is important, the application of our modelling results extends well beyond its confirmation of natural history. For a macro-scale example, fish owls are endangered within their range, which is presumed to be very large in Russia (Slaght & Surmach 2008), and there are many areas within the owl's range (including riparian areas within valleys) where the bird is not known to exist. The synoptic model provides the structure for a quantitative assessment of the entire Russian Far East relative to amount of potential habitat and the likely number of birds found there (home range sizes plus habitat conditions). In turn, the model can be used to predict the occurrence of fish owls within specific habitat conditions, which then can be tested with occupancy surveys for examining distributional limits, potential reintroduction sites and identification of critical watersheds (river systems that likely harbour multiple pairs of owls). The model also provides a structure for identifying patches within areas slated for logging or other development, which would aid conservation before or after proposed projects (see Slaght 2011). For a micro-scale example, Slaght (2011) and Slaght, Surmach & Gutiérrez (2013) showed that application of the synoptic model results could guide development of specific, regional recommendations within logging leases. Such leases account for 39% of potential fish owl habitat within our study area, whereas only 21% is protected. Logging companies are currently under no obligation to manage their leases for fish owls because there have been no defensible (i.e. quantitative) data on the owl in the region. We recommended buffer zones, roads management and limited harvest of old-growth trees in areas with high probability of use by fish owls; all based on predictions of the synoptic model (Slaght 2011; Slaght, Surmach & Gutiérrez 2013). Such recommendations have substantial implications, both economical (Furniss, Roelofs & Yee 1991; Thomas et al. 2006) and political (i.e. restricted-use areas based on modelling fish owl data are not an arbitrary construct, and therefore, cannot be easily dismissed by government officials).

Other Species

This model potentially solves a long-standing problem in estimation of home range size for some species, so there are many immediate applications of the synoptic model for conservation of species whose home ranges are constrained by environmental features. For example, fish space use and home range configuration could aid in predicting impacts of dams and guide stream habitat improvement, and conservation of riparian-dependent species would be greatly enhanced by accurate knowledge of their home range requirements and space use. Home range and resource use by species that use canyons or occur along oceanic shelves and beaches all would benefit from analysis using the synoptic model.

In addition, as home range estimates are often used to inform management decisions, the application of the synoptic model to species not restricted to linear habitat may be beneficial as well. For example, kernel density estimates of home ranges of Amur tigers Panthera tigris altaica living near the Sea of Japan coast in Russia (Goodrich et al. 2010) included large areas of marine habitat within tiger home ranges. This result was nonsensical, so the authors manually truncated the home range to approximate reality, which was necessary and logical as ocean is clearly not tiger habitat. In cases with other species, nonhabitat may not be as easy for researchers to identify, which can result in erroneous habitat inclusion or arbitrary truncation (which is not defensible to decision-makers). This has serious implications for conservation and management because it promotes an incorrect understanding of species’ biological requirements and undermines the credibility of scientists. By including habitat covariates such as ‘ocean’ in the synoptic model, nonhabitat can be identified objectively and excluded within the analytical process. Therefore, results will be more realistic, more credible and more effective in guiding conservation and management recommendations.

Acknowledgements

We thank S. Avdeyuk and other field assistants 2006-2010. T. Arnold, M. Bauer and D. Miquelle provided constructive input on this study. G. Hayward provided useful advice on study development and analysis. K. Slaght, D. Tempel, L. Berkeley, M. Kouffeld, C. Phillips, A. Barlow and three anonymous reviewers provided useful comments. One anonymous reviewer was especially helpful in his or her critiques of our paper. We are deeply grateful to D.S. Johnson for assessing the validity of model structure and interpretation. Financial and logistical support were provided by the Wildlife Conservation Society Russia program, United States Forest Service International Programs (grants 06-DG-11132726-215, 07-DG-11132792-153), Amur-Ussuri Centre for Avian Biodiversity, Disney Worldwide Conservation Fund, National Birds of Prey Trust, Bell Museum of Natural History, Columbus Zoo, Denver Zoo, Institute of Biology and Soil Sciences (Russian Academy of Sciences Far Eastern Branch), Minnesota Zoo, National Aviary, University of Minnesota (UMN), the Minnesota Agriculture Experiment Station, UMN Doctoral Dissertation Fellowship, the Gordon Gullion Scholarship and the Leigh H. Perkins Fellowship. This study was approved by UMN's Institutional Animal Care and Use Committee (Study Protocol Number 0610A95546).

Ancillary