Table 4. Analysis of covariance to test the stability ofthe spread parameter the power law model for dispersal gradients based ontemporal and spatial regression of the spread of cucurbit downy mildew in the eastern United States.
| Model used to estimate parameterb | ||||||||
|---|---|---|---|---|---|---|---|---|
| Temporal regression | Spatial regression | |||||||
| Analysisa | Source | DF | F-value | Pr > F | Source | DF | F-value | Pr > F |
| I | Year (Y) | 6 | 6.02 | 0.0001 | Year (Y) | 7 | 1.91 | 0.1274 |
| Time (T)b | 1 | 186.25 | 0.0001 | Distance (D) | 1 | 35.01 | 0.0001 | |
| Y × T | 6 | 4.54 | 0.0005 | Y × D | 6 | 6.12 | 0.0012 | |
| II | Year (Y) | 3 | 16.71 | 0.0001 | Year (Y) | 5 | 0.45 | 0.8021 |
| Time (T)b | 1 | 169.63 | 0.0001 | Distance (D) | 1 | 59.92 | 0.0001 | |
| Y × T | 3 | 9.36 | 0.0001 | Y × D | 4 | 3.09 | 0.0522 | |
Notes.
a
Analysis I is based on disease data collected in all epidemic years. Analysis II is based on data in years where the data were well described by the power law model. The model best described the data in 2008, 2009, 2010 and 2013 for the temporal regression, and in 2008, 2009, 2011, 2012 and 2013 for the spatial regression (see Table 2).
b
Time is measured on a monthly scale, i.e., in 4-week intervals.