Recent and future sea surface temperature trends in tropical pacific warm pool and cold tongue regions

Abstract

Using coral data, sea surface temperature (SST) reanalysis data, and Climate Model Intercomparison Project III (CMIP3) data, we analyze 20th-century and future warm pool and cold tongue SST trends. For the last 100 years, a broad La Nina-like SST trend, in which the warming trend of the warm pool SST is greater than that of the cold tongue SST, has appeared in reanalysis SST data sets, 20C scenario experiments of the CMIP3 data and less significantly in coral records. However, most Coupled General Circulation Models subjected to scenarios of future high greenhouse gas concentrations produce larger SST warming trends in cold tongues than in warm pools, resembling El Nino-like SST patterns. In other words, warmer tropical climate conditions correspond to stronger El Nino-like response. Heat budget analyses further verify that warmer tropical climates diminish the role of the ocean’s dynamic thermostat, which currently regulates cold tongue temperatures. Therefore, the thermodynamic thermostat, whose efficiency depends on the mean temperature, becomes the main regulator (particularly via evaporative cooling) of both warm pool and cold tongue temperatures in future warm climate conditions. Thus, the warming tendency of the cold tongue SST may lead that of the warm pool SST in near future.

Appendix: Meaning of the slope in this study

Approximating the linear trends of variables X and Y using the method of least-squares gives the equations

$$ X = a^{\prime } t + b^{\prime } $$

(1)

$$ Y = a^{\prime \prime } t + b^{\prime \prime } $$

(2)

Then, the equation for the least-squares line between X and Y becomes

$$ Y = aX + b, $$

(3)

where the constants a (slope) and b (intercept) can be found using the following equations:

$$ a = \frac{{N\sum XY - \left( {\sum X} \right)\left( {\sum Y} \right)}}{{N\sum X^{2} - \left( {\sum X} \right)^{2} }} $$

(4)

$$ b = \frac{{\left( {\sum Y} \right)\left( {\sum X^{2} } \right) - \left( {\sum X} \right)\left( {\sum XY} \right)}}{{N\sum X^{2} - \left( {\sum X} \right)^{2} }} $$

(5)

where N is the number of data. After substituting (1) and (2) into (4) and (5), we have a = a″/a′ and b = b″ − b′(a″/a′) so that a in Eq. (3) is the ratio between the trend of Y and that of X. Therefore, if a is greater than one, then the linear trend of Y is greater than that of X, and vice versa. However, this condition is only valid when a′ and a″ are the same sign, and otherwise a has a negative value indicating that a′ and a″ have opposite linear trends.