Environmental limits to growth: physiological niche boundaries of corals along turbidity–light gradients

Abstract

The physiological responses of organisms to resources and environmental conditions are important determinants of niche boundaries. In previous work, functional relationships between organism energetics and environment have been limited to energy intakes. However, energetic costs of maintenance may also depend on the supply of resources. In many mixotrophic organisms, two such resource types are light and particle concentration (turbidity). Using two coral species with contrasting abundances along light and turbidity gradients (Acropora valida and Turbinaria mesenterina), we incorporate the dual resource-stressor roles of these variables by calibrating functional responses of energy costs (respiration and loss of organic carbon) as well as energy intake (photosynthesis and particle feeding). This allows us to characterize physiological niche boundaries along light and turbidity gradients, identify species-specific differences in these boundaries, and assess the sensitivity of these differences to interspecific differences in particular functional response parameters. The turbidity-light niche of T. mesenterina was substantially larger than that of A. valida, consistent with its broader ecological distribution. As expected, the responses of photosynthesis, heterotrophic capacity, respiration, and organic carbon loss to light and turbidity varied between species. Niche boundaries were highly sensitive to the functional responses of energy costs to light and turbidity. Moreover, the study species’ niche differences were almost entirely attributable to species-specific differences in one functional response: that of respiration to turbidity. These results demonstrate that functional responses of energy-loss processes are important determinants of species-specific physiological limits to growth, and thereby of niche differences in reef corals. Given that many resources can stress organisms when supply rates are high, we propose that the functional responses of energy losses will prove to be important determinants of niche differences in other systems as well.

Appendices

Appendix 1

Rate of photosynthesis versus average daily irradiance

To test the efficacy of using average irradiance during the day rather than irradiance-time profiles in our calculations of daily rates of photosynthesis, we compared the outputs of these two methods for 12 days with varying maximum (1200 hours) irradiances (I_noon, 10–1,500 μmol m⁻² s⁻¹). Method 1 used a traditional irradiance sine function with a 12-h day length as input [I_t=I_noon sin(tπ/12), where I_t is irradiance averaged over a 1-h window and t is hours since sunrise], whereas method 2 used the average daily irradiance (I) as input. Daily gross rate of photosynthesis for method 1 was calculated as

$$P_{{\text{g1}}} = \sum\limits_{t = 0}^{12} {P_{\max } \tanh \frac{{I_t }}{{I_k }}} ,$$

(10)

and for method 2 as

$$P_{{\text{g}}2} = 12P_{\max } \tanh \frac{I}{{I_k }},$$

(11)

where P_max is maximum hourly rate of photosynthesis and I_k is the irradiance at which the rate of photosynthesis is ~75% of maximum. P_max and I_k are functions of I according to Eq. 8 (Anthony and Hoegh-Guldberg 2003a). Plotting P_g2 against P_g1 indicates negligible bias (P_g2≅ P_g1) for 1200 hours irradiances below 300 μmol m⁻² s⁻¹ (Fig. 3) . Thus, for the low light regimes characteristic of high-turbidity environments, the two methods produce nearly identical estimates of daily rates of photosynthesis.

Fig. 3

Plotting of P_g2 against P_g1 indicates negligible bias from the unity line

Appendix 2

Confidence limits of niche boundaries using Monte Carlo analysis

To model energy balance for a given turbidity and irradiance, we used parameter estimates, their statistical variances, and (where applicable) their statistical covariances. We followed standard Monte Carlo procedure, sampling sets of parameter values from the multivariate normal distribution specified by the parameter estimates and their associated variance covariance matrices (Table 5). This procedure was repeated for a range of turbidity (1–200 mg l⁻¹) and irradiance (0–300 μmol m⁻² s⁻¹) values. The sampling procedure was repeated 1,000 times for each turbidity-irradiance combination, and the standard deviation at each combination used as the confidence limits for the location of zero-E_B isoclines.

Table 5 Variance–covariance matrices for photosynthesis and respiration (A) and organic carbon loss parameters (B) for A. valida and T. mesenterina. See Table 4 for a summary of parameter means. Note that variances and covariances for β₁ and β₂ (the acclimation responses of P_max and I_k to changing growth irradiances) were unavailable, so we used the mean values presented in Table 4 for those parameters. Also, because most of the parameters for heterotrophy (γ_{sat, ε τ} and c_org) were fixed at (high) values, variation in heterotrophy was simulated using only the variance in maximum feeding rate, F_max