Table 3.

Land cover profiles for high overlap areas, and mean elevation for the four annual periods

Annual period		Proportion of high overlap area classified as:				Mean elevation (metres)
		Built up & bare ground	Crops & open savannah	Seasonally flooding grassland	Woodland & dense savannah (2-tailed)
Dec 1998 - Nov 1999	Cases	10.48%	61.05%	10.09%	18.38%	1098.27
	Controls	10.19%	55.21%	12.15%	22.44%	1095.37
	Difference	0.29%	5.84%	−2.06%	−4.07%	2.90
	p-value	0.72	>0.99	>0.99	<0.001	>0.99
Dec 1999 - Nov 2000	Cases	6.96%	48.86%	26.38%	17.13%	1066.31
	Controls	4.54%	59.61%	11.04%	24.81%	1100.10
	Difference	2.42%	−10.76%	15.34%	−7.67%	−33.79
	p-value	>0.99	<0.001	<0.001	<0.001	<0.001
Dec 2000 - Nov 2001	Cases	1.54%	52.66%	17.14%	28.66%	1058.35
	Controls	7.88%	59.56%	9.53%	23.03%	1099.74
	Difference	−6.34%	−6.90%	7.61%	5.64%	−41.39
	p-value	<0.001	<0.001	<0.001	<0.001	<0.001
Dec 2001 - Nov 2002	Cases	0.90%	33.68%	41.60%	14.06%	1047.82
	Controls	5.81%	60.84%	10.02%	23.34%	1093.86
	Difference	−4.91%	−27.15%	31.58%	−9.27%	−46.04
	p-value	<0.001	<0.001	<0.001	<0.001	<0.001

P-values shown represent the significance of the difference in proportion (for land cover classes) from a z-test, or difference in mean (for elevation) from a t-test. The alternative hypotheses for the proportion of built up and bare ground and crops and open savannah classes and mean elevation were that the proportion or mean for case high overlap areas would be lower than for control high overlap areas (one-tailed test). The alternative hypothesis for the seasonally flooding grassland class was that the proportion would be higher for case high overlap areas than control high overlap areas (one-tailed test). The alternative hypothesis for the woodland and dense savannah class was that the difference in proportions would not be zero (two-tailed test).