Season affects fire behavior in annually burned humid savanna of West Africa

Fire Ecology

Table 2 Prediction of fire behavior variables from weather and fuel variables

Response and explanatory variables	s.e.e.	R²	F	P-value
Log(I) = 10.61 – 0.0589 FM^*** – 0.0187 AH.	0.5701	0.55	14.89	< 0.0001
Log(r) = 0.8036 – 0.0717 FM^*** – 0.0191 AH^*	0.5697	0.63	20.54	< 0.0001
H = 4.4963 – 0.0388 FM^* – 0.0189 AH^	0.3587	0.64	20.97	< 0.0001
T = 1028.161 – 73.324 GH – 4.226 FM – 3.466 AH^*	105.967	0.45	6.41	0.002
Rt = 0.8694 – 0.7908 FL^** + 0.0206 FM.+ 0.0221 AH^** + 0.02496 AT	0.7276	0.51	5.74	0.002
C = 106.755 – 6.452 GH^*	3.197	0.34	12.99	0.001

Results show best models found by stepwise backward multiple regressions with plot and year as crossed random effects (using the R step function of the lmerTest package). Response (predicted) variables: fire intensity (I), rate of fire spread (r), flame height (H), fire maximum temperature (T), fire residence time above 60 °C (Rt), and combustion efficiency (C). Explanatory (predictors) variables: fuel load (FL), fuel moisture content (FM), grass height (GH), wind speed (W), air humidity (AH), air temperature (AT). s.e.e. = standard error of estimation; R² = coefficient of determination; F = F value
^***P < 0.001
^** P < 0.01
^*P < 0.05
.P < 0.1
P > 0.1