Water Exchange and Circulation in Selected Kenyan Creeks.

Abstract

Tides, currents, salinities and temperatures were studied from 1995-1998 in three selected creeks on the Kenya coast (Gazi Bay, Tudor and Kilifi Creeks) in order to improve knowledge on circulation and water exchange between the creeks and the ocean. Locally available meteorological data, tide gauge data and historical cruise data were also analysed. A meteorological mast was used for detailed studies of sea surface heat fluxes. The studies were carried out focussing on the monsoon seasons; the Inter-Monsoon Long Rain (IMLR; April-May), the South East Monsoon (SEM; June-September), the Inter-Monsoon Short Rain (IMSR; October-November) and the North East Monsoon (NEM; December-March). Gazi Bay is open to the sea whereas Tudor and Kilifi Creeks have sheltered bays inside 2 deep and relatively narrow inlets. The creeks are similar in size (Area;::;15 106m ; volum~60 1Q6m\ Spring tidal ranges at the entrances were estimated at 3.01, 3.16 and 3.3m, respectively, with strong M2 dominance and form factors of about 0.2. The creeks are not restricted (sensu Kjerfve), but there are tidal asymmetries with ebb dominance in Gazi Bay and udor Creek. However, the innermost parts of Gazi Bay feature flood dominance with strongly restricted flows and long ebbing periods (78hrs during spring). The features are accompanied by large shallow water and fortnightly tides. Similar shallow water and fortnightly response is seen also in the upper parts of Tudor Creek but the asymmetry is of different origin. Here, the tide is ebb dominated, presumably because of baroclinic processes at the inlet, and the tidal range is larger (5%) than in the ocean. In Gazi Bay the range is lowered by 15%. The asymmetries in Gazi Bay and Tudor Creek are accompanied by phase lags of 15-120min, indicating progressive character of the tide. Kilifi Creek data indicated a pure standing wave character. Kilifi inlet is shorter and wider than that of Tudor Creek, which probably explains the difference. Tide gauge temperature data from all sites are evaluated using harmonic analysis. Clear response for semidiurnal, diurnal and fortnightly periods indicates possibilities to unprove calculations of water exchange. Water exchange between the creeks and the ocean wa determined from a seasonal, steady state heat budget using temperature data from hydrographic observations and net sea surface heat fluxes from climatology data. In Tudor Creek th residence time is from 3 days at spring to 7 days at neap and in Kilifi Creek from 2.5 days at spring to 6 days at neap. In Gazi Bay the corresponding spring and neap residence times are 1 and 2.5 days, respectively. Corresponding calculations based on salinity basically failed because we were unable to determine the river discharge with a good enough resolution. Temperature and salinity differences between the creeks and the open ocean, averaging ~T;::;l °c, ~S::::-O.3, were used study the impact of the coastal ocean waters on the creek hydrography. These studies revealed interesting features, some of which are not foood in literature. There are two outstanding seasons, namely the late NEM (during February) and the IMLR (April/May) in between which rapid changes take place in the coastal waters. February is in the end of the dry season, dominated by excess evaporation and hypersaline conditions, May by strong rains estuarine conditions. In the ocean February is characterized by a low oceanic sea level, and a weak or occasionally nonexistent north-going EACC. The south-going Somali Current with high salinities may appear at the coast (even subducted). When the NEM weakens the situation changes; there is a rapid increase in the sea level and a decrease in salinity caused by rainfall at the onset of the IMLR We show a strong correlation between the steric height and seasonal development of the sea level. However, the interpretation suffers from too few ocean data. There is no clear whether the different seasons exert different effects on the water exchange. We expected the IMLR season to be more efficient in removing creek water, thus enhancing water exchange in comparison with the NEM because ofhigh current velocities. It seems as if this is the case, although the difference is small.