Calculation of the Bowen ratio in the tropical Pacific using sea surface temperature data

Abstract

The equilibrium Bowen ratio (Bo*) can be estimated empirically from the sea surface temperature (SST) only when the air is saturated with water vapor at the sea surface. However, since most of the sea surface is not saturated with water vapor, several studies have been conducted for the empirical Bowen ratio (BoE = Bo*(SST)), which was formulated with many sources of data from many areas. Hicks and Hess' [1977] linear regression formula for the BoE as a function of SST alone was used to estimate the empirical Bowen ratio in the tropical Pacific. Climatological and long‐term time series Optimum Interpolation Sea Surface Temperature (OISST) data was used for the climatological study of the BoE. The climatological mean Bowen ratio showed the regions, where the latent heat flux was most dominant. The Bowen ratio in the western Pacific warm pool (WPWP) area (O (0.04)) was 10% smaller than those in the upwelling coastal areas (O (0.14)). Along 5°N and 5°S, a 0.7 correlation coefficient between the BoE and the bulk formulated Bowen ratio (BoB), using climatological data, was found in the weak wind zone of the doldrums. BoB is the ratio of the bulk formulated sensible heat flux to the bulk formulated latent heat flux. EOF analysis was used to find out dominant temporal and spatial signals of the BoE. The first temporal EOF of the BoE, using 156 months (1986–1999) of OISST anomaly was related to El Niño events with an approximate 12‐month lag when compared to the El Niño‐Southern Oscillation (ENSO) index (SOI). The second and third temporal EOF showed annual and interannual signals, respectively. In order to estimate the discrepancy between the BoE and the BoB and to correct for the wind effect and the humidity effect, long‐term time series of Tropical Atmosphere Ocean (TAO) data was used. Our results showed that the eastern Pacific demonstrated a stronger correlation between the BoB and the BoE than the western Pacific. However, the Bowen ratio differences (BoD = |BoE − BoB|) were small in the western Pacific, and a larger margin of error was found at the eastern Pacific, where ocean dynamics perturb the upper ocean heat balance. The sensitivity of the humidity and the winds on the BoE was tested using linear correlation of the BoD while varying the winds and humidity. The BoD was small, especially in the southwest Pacific, where there are weak winds and high levels of the humidity. Thus the BoE in our study will be more useful to study involving heat fluxes in the WPWP area rather than those in the eastern Pacific. The correction of the humidity and the wind was applied using a linear regression of the humidity and the winds. After the corrections, the BoD was reduced by an average of 70.2%. Using TAO data, we obtained BoE = 0.033Bo* in the western Pacific with Bo*'s standard deviation of 0.008 at sea surface temperature 29.27°C ± 0.54°C.