Do yearly temperature cycles reduce species richness? Insights from calanoid copepods

Abstract

The metabolic theory of ecology (MTE) has explained the taxonomic richness of ectothermic species as an inverse function of habitat mean temperature. Extending this theory, we show that yearly temperature cycles reduce metabolic rates of taxa having short generation times. This reduction is due to Jensen’s inequality, which results from a nonlinear dependency of metabolic rate of organisms on temperature. It leads to a prediction that relatively lower species richness is found in habitats with larger amplitudes of yearly temperature cycles where mean temperatures and other conditions are similar. We show that metabolically driven generation time of a taxon also relates functionally to species richness, and similarly, its yearly cycles reduce richness. We test these hypotheses on marine calanoid copepods with 46,377 records of data collected by scientific cruise surveys in Mediterranean regions, across which the temperature amplitudes vary dramatically. We test both bio-energetic and phenomenological effects of temperature cycles on richness in 86 1° × 1° latitudinal and longitudinal spatial units. The models incorporated the effect of both periodic fluctuations and mean temperature explained 21.6% more variation in the data, with lower AIC, compared to models incorporated only the mean temperature. The study also gives insight into the basis of energetic-equivalence rule in MTE determining richness, which can be governed by generation time of taxon. The results of this study lead to the proposition that amplitude of yearly temperature cycles may contribute to both the longitudinal and the latitudinal differences in species richness and show how the metabolic theory can explain macro-ecological patterns arising from yearly temperature cycles.

Appendix 1: Calibrating generation timesg T

We calibrated the mean calanoid generation time per species, g _T , at temperature T (°C) using generation time data of copepods from Huntley and Lopez (1992) simplifying the model Eq. (8), ${g}_T={g}_0\overline{M^{1/4}}{\mathrm{exp}}^{E/ kT}$, as g _T  = ℏ ₁ exp[(ℏ ₂/1000)(1000/T)], with the selection of parameters, ${\hslash}_1={g}_0\overline{M^{1/4}}$, ℏ ₂ = E/k, and using the nonlinear least squares regression method lsqcurvefit in Matlab. The generation time data are given for 181 estimates in Huntley and Lopez (1992) of 33 marine copepod species tested at environmental temperatures ranging from − 1.7 to 30.7 °C. In our model calibrations, we should get ℏ ₂/1000 ~ 9K, ideally, as the activation energy E for aquatic taxa is ~ 0.78 eV (Allen et al. 2002). For our study of calanoid, we assumed that g _T calibrated based on marine copepod data, yielding R ² = 0.88 (Fig. 5a), are not deviated largely from that of calanoid copepods, which usually comprises of 55–95% of the plankton samples (see Mauchline and Mauchline 1998). We assumed also that size variation among calanoid species is negligible (most are 0.5–2.0 mm in length: Mauchline and Mauchline [1998]), and the distribution is Gaussian. As it is said that the generation time is a reasonable period for acclimatization of copepods in temperature-related experiments (Landry 1975; Huntley and Lopez 1992), we assumed that metabolism, and, thus, ${\overline{g}}_T$, responds to temperatures averaged at the generation time or a larger time-scale. As the generation time of marine copepods in the sampling units of the Mediterranean and adjacent seas, computed using the calibrated model as above, on the basis of the mean temperatures of the units, was less than 30 days (Fig. 5b), the assumption that monthly mean temperatures of the habitats is a reasonable scale, at which the copepods respond metabolically to varying temperature, may be justified.