Analysis Level One: Location
For the first level of analysis, the distribution of data was fairly normal with comparable variation between sample populations for each activity type (Figure 3). The control treatment (without light) saw a slight positive skew, particularly in vocalizations (Figure 4). There was more variation in the detection rate of Common Nighthawk vocalizations than wing booms (Figures 5 and 6). |
Analysis Level Two: Location x Sun Class
The control (dark) habitats saw a slight positive skew in observed vocalization detection rates throughout the active sun periods, while the treatment (light) habitats saw a negative skew during civil twilight (Figure 7). Typically, greater variation in the observed vocalization detection rates occurred at stations with artificial light (Figure 8). Overall, the greatest variation in the detection of Common Nighthawk vocalization rates occurred during the nautical twilight sun class periods in either sample population. The variation within and between treatment groups contributed to unequal variances, evident in the residual plot (Figure 9). These variances saw some correction with an inverse transformation (Figure 10). Slight skew is also observable in the residual plots. |
|
|
Likewise, wing boom detection rates were positively skewed in habitats without artificial light (Figure 11). Converse to the vocalization rates, wing boom detection rates tended to be negatively skewed at stations with artificial light except during civil twilight (Figure 12). There was less variation in wing boom detection rates between treatment groups compared to the vocalization detection rates. However, unequal variances persisted in the residual plot even when an inverse transformation was applied (Figure 13). |