![Page 1: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/1.jpg)
Effects of landscape pattern on the distribution of coastal cutthroat trout in headwater catchments of western Oregon
U.S. Department of the InteriorU.S. Geological Survey
Christian E. Torgersen1, Robert E. Gresswell2,Douglas S. Bateman3 and David P. Hockman-Wert1
1 USGS Forest and Rangeland Ecosystem Science Center, Corvallis2 USGS Northern Rocky Mountain Science Center, Bozeman, MT3 Dept. of Forest Science, Oregon State University, Corvallis
![Page 2: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/2.jpg)
![Page 3: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/3.jpg)
Landscape ecology of coastal cutthroat troutLandscape ecology of coastal cutthroat trout
Timber harvestTimber harvest
Climate, geology, and topographyClimate, geology, and topography
Natural and Natural and anthropogenic anthropogenic disturbancedisturbance
![Page 4: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/4.jpg)
Spatial Spatial scalescale
1:24,000 streams1:24,000 streamsUSGS 7.5USGS 7.5--min. quad.min. quad.
Population dynamics, Population dynamics, patch size, patch size, sampling/sampling/monitoring,monitoring,habitathabitat conservationconservation..
![Page 5: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/5.jpg)
Spatial Spatial extentextent
1:24,000 streams1:24,000 streamsUSGS 7.5USGS 7.5--min. quad.min. quad.
FishFish--bearing streamsbearing streams
![Page 6: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/6.jpg)
ObjectivesObjectivesExamine variation in the Examine variation in the spatial patternsspatial patternsof coastal cutthroat trout distribution of coastal cutthroat trout distribution within watershedswithin watersheds in western Oregon.in western Oregon.
Develop Develop explanatory modelsexplanatory models for predicting for predicting cutthroat trout distribution based on cutthroat trout distribution based on landscape characteristicslandscape characteristics..
Evaluate potential effects of Evaluate potential effects of landscape landscape disturbancedisturbance on cutthroat trout populations on cutthroat trout populations in headwater streams.in headwater streams.
![Page 7: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/7.jpg)
Site selectionSite selection
Isolated headwater Isolated headwater populations ofpopulations of
coastal cutthroat trout coastal cutthroat trout ((NN = 269)= 269)
Coast RangeCoast Range
CascadesCascades
KlamathKlamath
HardHard SoftSoftEcoregionEcoregionRock typeRock type
(500(500--1000 ha)1000 ha)
((GresswellGresswell et al. 2004. GIS/Spatial et al. 2004. GIS/Spatial Analyses in Fishery and Aquatic Sciences)Analyses in Fishery and Aquatic Sciences)
![Page 8: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/8.jpg)
(500(500--1000 ha)1000 ha) Site selectionSite selection
Randomly selected Randomly selected populations populations ((nn = 40)= 40)
Coast RangeCoast Range
CascadesCascades
KlamathKlamath
HardHard SoftSoftEcoregionEcoregionRock typeRock type
((GresswellGresswell et al. 2004. GIS/Spatial et al. 2004. GIS/Spatial Analyses in Fishery and Aquatic Sciences)Analyses in Fishery and Aquatic Sciences)
![Page 9: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/9.jpg)
00 1 km1 km
Age 1+ coastal Age 1+ coastal cutthroat trout cutthroat trout in pools and in pools and cascadescascades
SingleSingle--pass pass electrofishingelectrofishing surveyssurveys>200 km surveyed in 40 catchments>200 km surveyed in 40 catchments
(Torgersen et al. 2004. GIS/Spatial Analyses in Fishery and Aqua(Torgersen et al. 2004. GIS/Spatial Analyses in Fishery and Aquatic Sciences)tic Sciences)
![Page 10: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/10.jpg)
Spatial Spatial scalescale
1:24,000 streams1:24,000 streamsUSGS 7.5USGS 7.5--min. quad.min. quad.
Population dynamics, Population dynamics, patch size, patch size, sampling/sampling/monitoring,monitoring,habitathabitat conservationconservation..
![Page 11: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/11.jpg)
Racks Creek0.6 km
Tucca Creek 1 km
Distance (km)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
γ ( d
)
0
1
2
3
4
Distance (km)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
γ ( d
)0
1
2
3
4
Spatial scale (km)
Spatial scale (km)
Spatial scale of variation of trout countsSpatial scale of variation of trout counts
((GanioGanio, Torgersen, and , Torgersen, and GresswellGresswell. 2005. . 2005. Frontiers in Ecology and the Frontiers in Ecology and the EnvironmentEnvironment))
![Page 12: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/12.jpg)
Observed range (m)200 400 600 800 1000
Pre
dict
ed m
odel
rang
e (m
)
0
200
400
600
800
1000
1200
Predicting the spatial Predicting the spatial scalescale of of variation in CCT distributionvariation in CCT distribution
•• Weak rock (+)Weak rock (+)
•• Hard rock (Hard rock (--))
•• Mean distance Mean distance between between tributaries (+)tributaries (+)
Multiple linear regression
Adjusted R 2 = 0.78P
redi
cted
sca
le (m
)
Observed scale (m)
(after accounting (after accounting for maximum for maximum separation distance separation distance in basin)in basin)
![Page 13: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/13.jpg)
Spatial Spatial extentextent of coastal of coastal cutthroat trout distributioncutthroat trout distribution
Racks Cr. Racks Cr. –– Coast RangeCoast Range NevergoNevergo Cr. Cr. –– CascadesCascades
0.10 km/km21.65 km/km2
![Page 14: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/14.jpg)
Predictor variables in multiple Predictor variables in multiple linear regression modelslinear regression models
•• Ecoregion; geology (categorical)Ecoregion; geology (categorical)•• Percent resistant rock (basin area)Percent resistant rock (basin area)
•• Percent oldPercent old--growth (basin area)growth (basin area)•• Road density (km/km^2)Road density (km/km^2)•• Shape factor Shape factor (basin length^2 / basin area)(basin length^2 / basin area)
•• Mean stream slope (10Mean stream slope (10--m DEM)m DEM)
•• Percent private ownership (basin Percent private ownership (basin area)area)
•• Residual topography (10Residual topography (10--m DEM; m DEM; basin volume / surface area)basin volume / surface area)
•• Mean January temperatureMean January temperature•• Mean annual precipitationMean annual precipitation
•• Percent young forest (Cohen et al. Percent young forest (Cohen et al. 2002)2002)
* Variables listed in order of * Variables listed in order of ascending AIC for models ascending AIC for models predicting spatial extentpredicting spatial extent
Model selection with Model selection with Akaike’sAkaike’s Information Information Criterion (AIC)Criterion (AIC)
![Page 15: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/15.jpg)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
1:1Adjusted R2 = 0.73n = 40
Predicted spatial extent (km/km2)
Obs
erve
d sp
atia
l ext
ent (
km/k
m2 )
Predicting the Predicting the spatial extentspatial extent of trout of trout distribution based on landscape variablesdistribution based on landscape variables
- mean stream slope + mean annual precip. + percent young forest
![Page 16: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/16.jpg)
Model validationModel validation
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Random (n = 40)Validation (n = 8)
1:1
Predicted spatial extent (km/km2)
Obs
erve
d sp
atia
l ext
ent (
km/k
m2 )
Adjusted R2 = 0.73n = 40
![Page 17: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/17.jpg)
Percent young forestPercent young forestPercent young forest predicted spatial Percent young forest predicted spatial extent but wasextent but was not a goodnot a good predictor ofpredictor oftotal numberstotal numbers ((R R 22 = 0.32).= 0.32).
What does this suggest about What does this suggest about young young forest structureforest structure and its relationship with and its relationship with trout distribution? trout distribution?
What are some What are some potential mechanismspotential mechanisms by by which timber harvest influences fish which timber harvest influences fish distribution?distribution?
![Page 18: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/18.jpg)
Hydrologic response Hydrologic response (Jones and Post 2004)(Jones and Post 2004)
-50
0
50
100
150
200
250
300
1 2 3 4 5 6 7 8 9 10
Abs
olut
e ch
ange
(mm
) in
stre
amflo
wClearcutsClearcuts produce persistent spring surplusesproduce persistent spring surpluses
Time since disturbance (years)
conifer, seasonal snowconifer, transient snow
deciduous, transient snow
deciduous, seasonal snow
35302520151050 40
![Page 19: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/19.jpg)
Alternative hypothesesAlternative hypothesesMultiple hypothesesMultiple hypotheses explaining the explaining the association between the spatial extent of association between the spatial extent of trout distribution and young forest trout distribution and young forest structure.structure.
EffectsEffects of disturbanceof disturbance on vegetation, on vegetation, geomorphology, and temperature.geomorphology, and temperature.
Complex biological responsesComplex biological responses of trout to of trout to changes in cover, productivity, and changes in cover, productivity, and physical constraints to movement. physical constraints to movement.
![Page 20: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/20.jpg)
ConclusionsConclusionsLinks between Links between geology, geomorphologygeology, geomorphology, , and and fish distributionfish distribution
Relationships between Relationships between fish distributionfish distribution, , hydrology, hydrology, andand cumulative watershed cumulative watershed effectseffects? ?
Using Using digital elevation modelsdigital elevation models (DEM)(DEM)and and remote sensing remote sensing to predict fish to predict fish distribution at broad scalesdistribution at broad scales
![Page 21: Effects of landscape pattern on the distribution of ... · Predictor variables in multiple linear regression models • Ecoregion; geology (categorical) • Percent resistant rock](https://reader035.vdocuments.net/reader035/viewer/2022063003/5f5cded5825362637d35311a/html5/thumbnails/21.jpg)
Acknowledgments