Description of Data
Weighing pika in the field
Since 1995, Dr. David Hik and his research crew from the University of Alberta, have studied alpine herbivores (collared pikas, hoary marmots and arctic ground squirrels) at Pika Camp and their interaction with vegetation and other environmental factors. A particular focus has been the effects of changing climate on alpine vegetation and herbivory.
Pika trap
Over the last 15 years, the team has assembled a comprehensive dataset on the site`s collared pika population, including occupancy of talus patches and haypiles, demographics (sex and age), and measurements associated with their local environment (vegetative cover, characteristics of talus at haypile locations, haypile weight and content). Due to intensive trapping and survey efforts it is estimated that more than 95% of all individuals within the study area are captured annually (Morrison and Hik, 2007).
This project used data from the East and West sub-areas of the study site (Figure 4 on the Study Area page). Figure 5 below shows the number of pikas in the East and West between 1995 and 2005, the time period of this study. As seen in this graph, both sub-populations had declined almost to the point of complete extirpation in 2000. They rallied in 2002 and declined again in 2003. Numbers have been increasing since 2004.
This project used data from the East and West sub-areas of the study site (Figure 4 on the Study Area page). Figure 5 below shows the number of pikas in the East and West between 1995 and 2005, the time period of this study. As seen in this graph, both sub-populations had declined almost to the point of complete extirpation in 2000. They rallied in 2002 and declined again in 2003. Numbers have been increasing since 2004.
Figure 5. Patterns of patch occupancy by pikas 1995 - 2005
Spatial Data
Spatial data are measures of the area and perimeter of each talus patch and the connectivity between patches. Connectivity is a complicated calculation of the measure of connectivity:
(Si = Σj≠i exp(−α(dij))Aj^b) where dij is the distance between patches i and j, ą is the inverse of average migration distance, Aj is the area of patch(j), and b is a buffer measure with b=1 (Moilanen and Nieminen 2002). Put simply, the larger the connectivity index , the greater the effective connectivity between patches.
(Si = Σj≠i exp(−α(dij))Aj^b) where dij is the distance between patches i and j, ą is the inverse of average migration distance, Aj is the area of patch(j), and b is a buffer measure with b=1 (Moilanen and Nieminen 2002). Put simply, the larger the connectivity index , the greater the effective connectivity between patches.
Environmental Data
Sample haypile
Comprehensive data has been assembled for each of 72 haypile locations in the east and west sub-areas, including:
· Aspect and slope
· Talus measurements (e.g., boulder size, spacing,
whether a patch is level or concave)
· Distance from the patch to vegetation
· Interspersion of talus and meadow
· Composition of adjacent vegetation
These measurements are assumed to have remained constant over the eleven-year study period.
A list of spatial and environmental variables and their analysis codes is shown in Table 1.
· Aspect and slope
· Talus measurements (e.g., boulder size, spacing,
whether a patch is level or concave)
· Distance from the patch to vegetation
· Interspersion of talus and meadow
· Composition of adjacent vegetation
These measurements are assumed to have remained constant over the eleven-year study period.
A list of spatial and environmental variables and their analysis codes is shown in Table 1.
Table 1. List of spatial and environmental variables and associated analysis codes
Climate Data
Unseasonable snowy summer's day at Pika Camp
The following regional climate datasets were included in this analysis. A full list of climate variables and associated analysis codes is shown in Table 2.
a. Pacific Decadal Oscillation
The Pacific Decadal Oscillation (PDO) is a repeating pattern of variability in sea surface temperatures across the north Pacific Ocean, which has an effect on winter weather patterns throughout northwestern North America. Warm phase PDOs result in above-average temperatures and below-average winter precipitation and snowpack conditions. Cold phase PDOs have an opposite effect. Morrison and Hik (2007) have documented a correlation between survival in adult pikas and large scale climate patterns (PDO) and associated timing of spring snowmelt. I will use PDO as a surrogate for the timing of snowmelt
(source of data: http://www.cdc.noaa.gov/Climate-Indices/Analysis).
b. Regional climate data
Data on seasonal temperature and precipitation were sourced through ClimateWNA v4.51 for the years 1994 – 2006. (source of data: http://www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html). Climate data includes average, minimum and maximum temperature and average precipitation for each of the four seasons, precipitation as snow, mean coldest and warmest month temperature and frost-free period. Using data collated by season rather than by calendar year makes it possible to better assess the potential effects of each winter's climate on pika population data collected the following summer.
Climate WNA data is based on calculated and derived climate values and is corrected for elevation.
a. Pacific Decadal Oscillation
The Pacific Decadal Oscillation (PDO) is a repeating pattern of variability in sea surface temperatures across the north Pacific Ocean, which has an effect on winter weather patterns throughout northwestern North America. Warm phase PDOs result in above-average temperatures and below-average winter precipitation and snowpack conditions. Cold phase PDOs have an opposite effect. Morrison and Hik (2007) have documented a correlation between survival in adult pikas and large scale climate patterns (PDO) and associated timing of spring snowmelt. I will use PDO as a surrogate for the timing of snowmelt
(source of data: http://www.cdc.noaa.gov/Climate-Indices/Analysis).
b. Regional climate data
Data on seasonal temperature and precipitation were sourced through ClimateWNA v4.51 for the years 1994 – 2006. (source of data: http://www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html). Climate data includes average, minimum and maximum temperature and average precipitation for each of the four seasons, precipitation as snow, mean coldest and warmest month temperature and frost-free period. Using data collated by season rather than by calendar year makes it possible to better assess the potential effects of each winter's climate on pika population data collected the following summer.
Climate WNA data is based on calculated and derived climate values and is corrected for elevation.
Table 2. Regional climate variables and associated analysis codes
Figures 6 and 7 below show the results of preliminary graphing of population variables as a function of climate measures. As can be seen from these graphs, there is evidence of a relationship between frost-free period and number of pikas/year (Figure 6) and maximum winter temperature and numbers of extirpations (Figure 7).
Figure 6. Number of pikas/yr as a function of frost-free period (Julian days)
Figure 7. Number of local extirpations as a function of maximum winter temperature
Table 3. Sample data table: patch scale analysis
A sample of patch scale data is shown in Table 3. Data is organized into columns that group response (population) variables and predictor (spatial and environmental) variables.
Correlations among data were checked with scatterplots and using a Pearson’s correlation test to highly correlated variables were removed. Figure 8 demonstrates a test of correlation among vegetation factors. Relationships between variables were also tested early on using scatterplots or line graphs, as appropriate (see climate examples in Figures 6 and 7 above).
Figure 8: Multipanel scatterplot to test for correlation between vegetation factors
