Does a high level of human activity and site type affect the level of coyote activity?

Statistical and Graphical Results

                   Figure 9. Effect of human relative activity on coyote relative activity in natural, rural and urban sites. The 
                   red line represents a logarithmic regression model curve fitted to the overall data from all site types.

                Table 3. Overall correlation between human relative activity and coyote relative activity across all site types.

                     Table 4. Correlation between human relative activity and coyote relative activity in each specific site type.

Figure 9 illustrates the relationship between human relative activity and coyote relative activity. Across all three sites, the overall trend in the data shows that an increase in human relative activity is associated with a decrease in coyote relative activity. Table 3 shows the overall correlation for all three site types together. The correlation value of -0.455 indicates a moderate negative relationship between human and coyote relative activity, confirmed by the low p-value (2.20E-16). The red line represents a logarithmic regression model curve. The logarithmic model was the best fit for the overall data set, indicating that a drop in coyote relative activity is not proportionate to the associated increase in human relative activity.

Figure 9 also shows the relationship between coyote and human relative activity at each site type. The grey data points represent the natural site type and show that human relative activity is consistently low or absent, while there is a broad range of coyote relative activity. The gold data points represent the rural site type and show that slight increases in human relative activity are associated with large decreases in coyote relative activity. The blue data points represent the urban site type and also show that increases in human relative activity are associated with decreases in coyote relative activity, though the decrease appears to be less significant than in the rural site.

Table 4 shows the correlation between human and coyote relative activity for each site type. Both the rural and urban sites have low p-values and correlation values indicating moderate negative correlations between human and coyote relative activity (more so in the rural sites than in the urban sites). The natural sites have a high p-value (0.1359) and a correlation value of -0.137, indicating a very weak correlation between coyote and human relative activity. This result was expected as human activity was minimal at the natural site and coyotes were active throughout 24-hour periods. 

Figure 10. Average coyote and human relative activity levels across a 24 hour period in natural, rural and urban site types.

Table 5. Two-way ANOVA results from coyote relative activity data across all three site types divided into diurnal and  nocturnal time periods.

Figure 10 shows the average coyote and human relative activity over a 24-hour period in the natural, rural and urban site types. As expected, coyotes in natural areas were active throughout the diurnal and nocturnal periods, while human activity was minimal or absent. In the rural and urban sites, coyote activity was lowest and human activity was highest from approximately 6 AM through 6 PM. Human activity was lowest and coyote activity was highest from 7 PM through 5 AM. The change in relative activity in urban coyotes over the 24-hour period appears more stable than the change seen in rural coyotes. Diurnal coyote relative activity in rural sites drops quite sharply from approximately 6 AM through 6 PM.

The data remained organized by site type, but was further subset by defining 6 AM through 6 PM as diurnal and 7 PM through 5 AM as nocturnal. A two-way ANOVA was then run to determine if any significant differences between diurnal and nocturnal activity existed between the site types. There were significant effects between site types and between times, but there was also a significant interaction effect between site types and times (Table 5). The high F-values and low p-values associated with the variables and interaction indicate the significant effects. The significant variable effects and interaction effect signaled a need to examine nocturnal coyote relative activity and diurnal coyote relative activity separately across the three site types using one-way ANOVAs, performing pair wise comparisons to determine where the significant differences were occurring.

Table 6. One-way ANOVA with pair wise comparisons results from coyote relative activity data for the diurnal period (6 AM through 6 PM) compared between each site type.

                          Figure 11. Average diurnal coyote and human relative activity in natural, rural, and urban site types.

Table 6 shows the results from the one-way ANOVA with pair wise comparisons for diurnal coyote relative activity across the three site types. The large F-value (174.56) and small p-value (< 0.001) indicate that there is a significant difference within the data. When urban diurnal coyote relative activity and rural diurnal coyote relative activity were compared to natural diurnal coyote relative activity, both comparisons showed a statistically significant difference (p-values of < 0.001). Surprisingly, when urban diurnal coyote relative activity was compared to rural diurnal coyote relative activity, the p-value was 0.9908, indicating that there was no difference in diurnal coyote activity between the two sites. Figure 8 shows a graphical representation of these findings, where the average coyote relative activity bars for the rural and urban site types are virtually identical. 

Table 7. One-way ANOVA with pair wise comparisons results from coyote relative activity data for the nocturnal period (7 PM through 5 AM ) compared between each site type.

                           Figure 12. Average nocturnal coyote and human relative activity in natural, rural and urban site types.

Table 7 shows the results for the one-way ANOVA with pair wise comparisons for nocturnal coyote relative activity across the 3 site types. Again, a large F-value (93.01) and a small p-value (< 0.001) indicate significant effects within the data. Urban nocturnal coyote relative activity was found to be significantly different from both rural and natural nocturnal coyote relative activity. Interestingly, rural nocturnal coyote relative activity was not significantly different from natural nocturnal relative activity. Coyotes in rural areas likely behave similarly to their natural counterparts during nocturnal periods as this is when human activity is lowest. Exploitation related activities take place primarily during the day, and as such rural coyotes have shifted their activity patterns to avoid any contact with humans in the diurnal period.

The urban coyote relative activity curve in figure 10 shows that coyote relative activity remains more constant throughout a 24 hour period, similar to natural coyotes but at a lower level. Urban coyote relative activity may be similar to rural coyote relative activity during the diurnal period, as management removes problem coyotes that are active in the daytime and as such have a higher risk of conflict with humans. Alternately, urban coyotes may be becoming habituated to human activity. An urban area cannot support larger coyote populations and this may explain the lower levels of activity. The consistency in activity levels is indicative that coyotes do not significantly alter their activity levels in response to increasing human activity levels during the day.