In general, it was found that temperature sensitivity and temperature standard deviation were the most important factors affecting the risk of extinction and time taken to go extinct. Species-richness showed a strong effect on mean extinction risk of species within a small product interval of temperature sensitivity and temperature variability. But connectance, Hill exponent and species preferences did not have any effects.

The results showed that:

• Extinction risk of species increased with increasing food web size.
• Food web robustness decreased with increasing food web size.
• Extinction risk of 50% of species decreased with decreasing food web size.
• Rapid increase in extinction risk when temperature variability crossed lower threshold.

Figure 4 shows the logistic regression curves with 95% confidence band [black, red, blue, green curve for 6, 12, 18, 24 species respectively] showing the risk of extinction as a function of the product of temperature sensitivity and temperature standard deviation. Formula denotes the linear regression equations. 10%, 50% and 90% extinction thresholds are also indicated in figure.

It was found that the time taken by the food web to lose 1/6 of the species decreased as the number of species in food web increased.

Figure 5 shows the average time taken by food web community to lose 1/6 of species from the web. Regression curves with 95% confidence band [(a) black curve for 6 species (b) red curve for 12 species web (c) blue curve for 18 species web and (d) green curve for 24 species] shows the average time taken to extinction as a function of the product of temperature sensitivity and temperature standard deviation. Formula denotes the linear regression equations with intercept and confidence interval around the slope coefficient.