This empirical model account- 

 ed for observed distributions 

 fairly well, implying a relatively 

 small contribution to S (x,t) by 

 higher order terms such as dis- 

 tance dependent S and other peri- 

 odic tidal components. The fit- 

 ted equations had high coefficients 

 of determination (1977 20 out of 22 

 had r^ > 0.88, 1978 17 out of 20 

 had r > 0.83) and estimated half- 

 tidal ranges were close to ranges 

 observed at MP -7 and MP 18 when 

 profiles were obtained at 1-3 h in- 

 tervals. Variations in k reflected 

 the decrease in longitudinal salinity 

 gradient from spring to summer and 

 stronger gradients in the upper than 

 the lower layer. Using the same 

 method on whole water column aver- 

 ages, an average lower estuary xross- 

 section area of 1.8 x 10 m , and 

 Q as calculated above, values of 

 E (= -Q f /kA) from 500 to 5000 m /s 

 were calculated over a Q range of 

 2 - 20 x 10 m d . Simpson and Ham- 

 mond (submitted) have suggested 500 

 to 2500 m /s for lower estuary over 

 a Q range of 2 - 10 x 10 m d 



VOLUME TRANSPORTS 



Mean transports under high 

 and low flow conditions were higher 

 than mean Q_ in each layer (Figure 



4). The ratio of Q to Q r ranged 



u 1 

 from 2 to 6 under high flow condi- 

 tions and from 5 to 12 under the 

 low flow conditions. Ratios of 

 10 to 40 have been reported in 

 other partially stratified estua- 

 ries, e.g. the James River (Prit- 

 chard 1967), Mersey Estuary 

 (Bowden 1960) , and Juan de Fuca 

 Straight (Tully 1958). Differ- 

 ences in this ratio between high 

 and low flow periods were primar- 

 ily due to changes in Q with Q 

 (and Q ) remaining relatively con- 



stant. Such stability over a wide 

 range of Q reflects the inverse 

 relationship between Q f and vertical 

 salinity gradients and is primarily a 

 consequence of an increase in verti- 

 cal exchange under low flow condi- 

 tions (Figure 4) . 



CHLOROPHYLL a DISTRIBUTIONS 



Mean upper and lower layer 

 Chlorophyll a concentrations for the 

 box model were computed from vertical 

 profiles of Chlorophyll a by averag- 

 ing over the depth ranges used to 

 average salinity. When profiles were 

 available from two consecutive days 

 at a given station they were aver- 

 aged. Tidal variations in Chloro- 

 phyll a were considerable but, unlike 

 salinity, tidally averaged profiles 

 were not calculated. Mean Chloro- 

 phyll a over two tidal cycles at sta- 

 tions MP -7 and MP 18 had average 

 coefficients of variation of 44 per- 

 cent (27 to 607o) in the upper and 31 

 percent (16 to 57%) in the lower 

 layer. However, the upper and lower 

 layer Chlorophyll a concentrations 

 were positively correlated (r 0.60 to 

 0.99, P< 0.05) except in May 1977, 

 1978 when r was not significant. This 

 implies that the between layer dif- 

 ferences used to calculate source and 

 sink terms were less variable than 

 Chlorophyll a concentrations used to 

 calculate upper and lower layer 

 fluxes . 



Flushing times were calculated 

 by applying Chlorophyll a distribu- 

 tion data to Eq. (7) and setting the 

 boundary inputs to zero. Solutions 

 for time varying Chlorophyll a ob- 

 tained through the method of similar- 

 ity transformations (Noble and Daniel 

 1977) had the general form of p, exp 

 (Ak 1 -) + p Q , where A, is the k 

 eigenvalue of A (Eq. 7) and the p's 



177 



