Performance of the SSI development function compared with 33 other functions applied to 79 arthropod species' datasets

The development rates of arthropods are temperature-dependent. Studies aiming to predict the dynamics of arachnid, crustacean, and insect populations in nature often require the derivation of development functions representing this phenomenon. A previous study (Quinn, B.K., 2017, J. Therm. Biol. 63, 65-77) identified 33 development functions commonly used in past studies on temperature-dependent development of arthropods, and illustrated that: (1) most of 99 past studies only applied one or few (2-5) development functions to their data without considering others; and (2) most of a subset of 79 studies' data were not fit with the actual best function for them, resulting in sometimes substantial differences in model performance and predictive ability. However, that study did not test the class of development functions based on theoretical enzyme thermodynamics, including the Sharpe-Schoolfield-Ikemoto (SSI) function. Herein, the meta-analyses done in that previous study were redone, after fitting all 79 reanalyzed datasets with the SSI function. Estimates of the intrinsic optimum temperature (T-Phi) for development of each tested species were also derived using the SSI function and compared among taxa. Including the SSI function in analyses did not change the conclusions of the previous study concerning development function usage, choice, and consequences. Notably, the SSI function performed as well as or relatively better than other functions of comparable or lower complexity in terms of R-2, AIC(C)-based rankings, Delta AIC(C) values, and prediction errors, which may recommend its more widespread use in future studies. Overall differences in T-Phi were found among arthropod subphyla, as well as between most species pairs. Most T-Phi estimates produced herein were novel, and could be used to make inferences about or comparisons among arthropod taxa in future studies.

Automated summary

The development rates of arthropods are temperature-dependent, and specific development functions are needed for research. Many past studies have only applied a few development functions and did not use the best-fit function for their data, suggesting the wider use of the SSI function in future studies.