Bias and Variability in Lower Developmental Thresholds Estimated from Field Studies

Abstract

We conducted simulated experiments to explore how sampling frequency could influence estimates of insect lower developmental thresholds made from field studies. We also investigated how a 9.44 or 18.89°C (17 and 34°F) difference between daily maximum and minimum temperatures and how 2 similar years, 2 dissimilar years, or 6 dissimilar years of temperature data could influence these estimates. Finally, we investigated how a -3.89 or 7.22°C (25 or 45°F) true lower developmental threshold and 2 different estimation methods, could influence estimates of insect lower developmental thresholds. Results indicated that having two very similar years of temperature data may greatly increase the variability of the estimated lower developmental thresholds. Also, sampling every 3, 4, or 7 d compared with sampling every day, can increase the amount of variability in the estimated lower developmental thresholds. Acceptable estimates of lower developmental thresholds may be obtained with as few as 2 yr of temperature data if the data came from 2 very dissimilar years (weatherwise). However, as many as 6 dissimilar years of temperature data may be needed if the true lower developmental threshold is approximately 7.22°C and the difference between daily maximum and minimum temperatures is 18.89°C. Almost all estimates for lower developmental thresholds were biased and tended to underestimate the true lower developmental threshold. Finally, we found no differences between the 2 estimation methods in bias, variability, and average estimates for the lower developmental thresholds.