5. HYDROGRAPHY

5.1 Introduction
5.2 Wind and freshwater outflow
5.3 Characteristic hydrographic features and water masses
5.4 Surface currents
5.5 Hydrographic structure
5.6 Discussion
5.7 Tables and figures

5.1 Introduction

The Mozambique Current is usually considered as a part of the anticyclonic subtropical gyre consisting of the South Equatorial Current, the Agulhas Current system and the eastward flow situated to the north of the subtropical convergence (WYRTKI, 1973). Several opinions exist concerning the status of the Mozambique Current. The classical concept of the Agulhas Current being an extension of the Mozambique Current has been questioned by MENACHÉ (1963). He claims that during at least some parts of the year the water transported by this current turns back at the southern mouth of the Mozambique Channel and flows north along the west coast of Madagascar. HARRIS (1972), however, observed that a fraction of the Mozambique Current passed through the southern part of the channel as one of the tributories to the Agulhas Current. LUTJEHARMS (1976), making use of all the available data from the north-east monsoon season, concluded that the origin of the inflow to the Agulhas Current is a function of depth and that the Mozambique Current is the major source for the upper layers during this season.

CREPON (1964) gives some indications of a looping of the West Madagascar Current to form the Mozambique Current. An interpretation of this observation together with that of MENACHÉ (1963) is that a closed anticyclonic circulation exists within the Mozambique Channel.

By isentropic analysis of data from spring 1964 HARRIS (1972) showed the presence of a front of constrasting water properties along the eastern boundary of the Mozambique Current from Durban to the northern end of the Mozambique Channel. Along this front a series of deep anticyclonic vortices were identified.

The core of the southward-moving Mozambique Current is generally found close to the continental slope. The northern part of the Mozambique Current as well as its main source, the South Equatorial Current, is directly influenced by the monsoon winds. Considerable seasonal variations in velocity and volume transported are therefore to be expected. The South Equatorial Current is strengthened during the South-West Monsoon in April-October but only one third of this water turns south as the East Madagascar and the Mozambique Current (WYRTKI, 1973). It is likely, however, that also this ratio is subject to seasonal variations.

According to the British Admiralty Sailing Directions the Mozambique Current has its minimum velocity during May-July and its maximum during August-January. DARBYSHIRE (1964) claims the Agulhas Current to be strongest in April and weakest in October. Other authors have shown that this current has its maximum speed both in spring and in autumn. The problem of the seasonal changes in the Mozambique and the Agulhas Currents has thus not yet been resolved (LUTJEHARMS 1977). PEARCE (1977) deals with the short-term variations of the Agulhas Current and has observed meanders over tens of kilometers off Durban.

Several authors have observed an inshore current along parts of the coast, flowing in a contrary direction to the main current. This counter current seems not to be a permanent feature but highly influenced by the prevailing winds (CLOVES, 1950). In general the counter currents are weak, but have occasionally been observed to flow at high speeds. STAVROPOULOS and DUNCAN (1974) present a current section off Durban showing maximum northward currents of 100 cm/s in the inner 20 km. In Beira Bay a strong and fairly permanent longshore counter - counter current have been reported (TINLEY, 1971).

5.2 Wind and freshwater outflow

Table 5.1 gives the monthly wind direction frequency at six localities (Fig. 2.2) along the coast of Mozambique from August 1977 to June 1978. The observations were only carried out during daytime at 0900, 1500 and 2100 hrs. The alternating monsoon system with winds from the north-east during the southern summer (November-April) and from the south-west during the southern winter (May-October) is pronounced along the northern coast south to Sofala Bay.

Along the southern part winds between North and East dominated from September to December both at Vilanculos and in Maputo. In August and from January to June winds between East and South were prevailing at Vilanculos while the situation was more variable in Maputo. Northerly winds were especially frequent in Maputo in May.

It is uncertain to which degree these coastal observations are representative for the more offshore areas. Some of the meteorological stations might be influenced by local topographic conditions. The coastal stations also show a diurnal variation in wind direction, with nocturnal offshore winds and onshore winds during daytime.

Fig. 6.1 shows the average monthly freshwater discharge from some of the main rivers of Mozambique. The localities of the water gauges appear in Fig. 2.2. As seen, there is a maximum freshwater outflow in February-March and a minimum in October. The whole Mozambican coast seems to belong to the same hydro-logical regime, with a possible exception for Maputo River which has a rather small seasonal variation. Unfortunately it has not been possible to obtain discharge data from one of the major freshwater sources, namely the Zambezi River. The lack of data makes it difficult to estimate the total freshwater outflow flow Mozambique. The average outflows shown in Fig. 6.1 are, however, believed to give a correct qualitative picture of the seasonal variations.

5.3 Characteristic hydrographic features and water masses

The hydrographic SECTION V from May 1978 shown in Fig. 5.2 elucidates some of the most conspicious hydrographic features found off the coast of Mozambique. A thermocline is observed usually at depths between 50 and 100 m. From this there is a gradual decrease in temperature with depth to about 2.5°C at about 2000 m. There is a subsurface salinity maximum at about 200 m depth. The vertical distribution of oxygen shows a minimum zone situated slightly above the salinity maximum. Another oxygen minimum is seen at 1000-1200 m. Between the shallow and the deep oxygen minima is situated an intermediate oxygen maximum. These features are found more or less pronounced in the whole Indian Ocean north of 40°S (WYRTKI, 1971).

Fig. 5.3 shows the temperature-salinity relationship from the same section as in the previous figure. This is representative for the whole Mozambican coast. Following mainly WYRTKI (1971, 1973) four main water masses can be identified:

Surface water. This is characterized by temperatures between 22°C and 30°C with salinity below 35.2‰. The surface water mostly has a salinity below 35.0‰ with increasing values southward. This water consists of the low-salinity tropical surface water transported westward across the Indian Ocean by the South Equatorial Current. The salinity of this water is kept low by the freshwater runoff from the coast of Mozambique.

Subtropical surface water. This water is formed in the center of the subtropical anticyclonic gyre where it sinks to more than 500 m. From here it spreads to form a sub-surface salinity maximum throughout the entire gyre system. Off the coast of Mozambique it is usually found at depths between 150 and 250 m. The highest salinity values of this maximum zone are observed along the southern part of the coast.

Antarctic intermediate water. This water mass is formed by mixing along the Antarctic polar front of the cold and low-salinity surface water of Antarctic and warmer water of higher salinity from the north. The mixing products sink and continue on north as the Antarctic intermediate water.

In Fig. 5.4 is plotted s_t versus temperature from the same section as the previous figures. The relationship could be rather nicely fitted to the following expressions:

st = 30.95 - 0.3 t	for t > 16°C
st = 28.10 - 0.12 t	for t < 16°C

These equations seem to be valid for the whole Mozambican coast except for surface water of salinity below 35.0‰. This means that the interior mass distribution of the hydro-graphic sections is fairly well described by the temperature field. For that reason no s_t sections are presented.

5.4 Surface currents

The surface currents were mapped by observing the drift of the vessel. Though the vessel was equipped with a satellite navigator allowing very precise determination of position, the time between each reliable satellite fix could be up to five hours. It is thus obvious that smaller details in the current picture, such as current shear and vortices, will not be observed. Additionally, the wind component effecting the drift of the vessel is difficult to estimate. In spite of these reservations, some conspicious features of the surface current are very pronounced in the drift observations:

In September 1977 there was a northward current off Cabo Delgado, and the dividing area of the South Equatorial Current seemed to be at about 11°20’S. During the rest of the observation period the southward-flowing Mozambique Current was observed along the whole coast. North of Nacala the strongest currents were usually found in the eastern part of the investigated area, about 80-100 km from the coast. From Nacala to Inhambane the highest current velocities were observed close to the continental slope with decreasing values eastwards. At the southern end of the Boa-Paz Bank the current core moves toward South-South-West without the previous tendency to follow the bathymetrical curves and without intruding into Delagoa Bay.

The maximum velocity observed was about 200 cm/s and the highest values were in general found north of 18°S. The area off Angoche seemed to be the area of maximum current velocity.

Northward counter currents were observed between 26° and 27°S, in Delagoa Bay and at Sofala Bank. These appeared to be most pronounced and reached the highest velocities, up to 100 cm/s, during February-March.

5.5 Hydrographic structure

September 1977. The horizontal distribution of temperature is shown in Fig. 5.5. The temperature decreased from 25.5°C off the northern part of the coast to about 23°C in Delagoa Bay. A tongue of slightly cooler water extended southward from the mouth of Zambezi River which was probably associated with the freshwater outflow from this river.

The highest salinities were found close to the coast along the northern part of the coast south to Quelimane (Fig. 5.6). Water of salinity below 35‰ was recorded 70-130 km off the coast. The outflow from Zambezi River could be seen as a tongue of low salinity water propagating southward. North-west of Inhambane more saline water was brought into the area from the east.

Figs. 5.7-5.12 give the vertical distributions of temperature, salinity and oxygen in the six hydrographic sections. In all the sections the characteristic features described in Ch. 5.3 of a salinity maximum at about 200 m and a oxygen minimum zone slightly above this maximum, are pronounced.

As an aid in the interpretation of the hydrographic data, the vertical geostrophic velocity distribution relative to the surface was calculated for each hydrographic section.

At SECTION I (Fig. 5.7) there seemed to be two cores of maximum southward velocity; one about 50 km from the coast and another at the outermost stations 140-150 km seawards. In the inner 30-40 km a northward current seemed to occur, which was associated with the lowsalinity water of the upper 100 m. The northward current seemed to have its maximum value close to the coast.

The maximum southward current in SECTION II (Fig. 5.8) was probably found about 50 km from the coast. In the part of the section nearest the coast indications of a weak northward current were seen. Centered about 100 km from the coast there seemed to be a cyclonic eddy. This eddy was most pronounced in the deep layer below the thermocline.

In SECTION III (Fig. 5.9) two maximum zones in the southward current were indicated, one 50-60 km from the coast and another 100-120 km seawards. There were also indications of a coastal counter current over the shelf.

In SECTION IV (Fig. 5.10) the tongue-like lowsalinity water from the Zambezi River which was pronounced in the surface salinity distribution (Fig. 5.6) was seen in a narrow zone over the shelf. The baroclinic structure of the section was weak and southward transport through the section seemed to be negligible. A maximum northward current seemed to occur 110-120 km offshore. It also seemed from the surface salinity distribution as the southward flow was deflected to the east and northeast, and that only a small fraction of this water passed further south.

It was also indicated by the surface salinity distribution that more saline water from the east was being fed into the Mozambique Current along two possible routes, one east of Bazaruto Island and another north-east of Inhambane.

The latter could be followed in SECTION V (Fig. 5.11) as a zone of decreased temperature and increased salinity. In this section the subtropical surface water usually found in about 200 m was lifted to the surface. This appeared to be due to a cyclonic eddy which was situated outside the southward long-shore flowing water.

The northbound current indicated at the inner 20-30 km of SECTION VI (Fig. 5.12) was also observed by the drift of the vessel. The southward current core seemed to be about 50 km offshore. Further seawards a cyclonic circulation was indicated, resulting in a pronounced shear zone clearly seen in all three parameters.

The depth to the thermocline, D, can be defined as t_O - t_D < 1°C, where t_O is the temperature at the surface and t_D the temperature at the upper boundary of the thermocline. The thickness of this mixed and homogeneous layer was about 100 m along the northern coast with decreasing values southwards and when approaching the coast.

November 1977. During this cruise the surface temperatures (Fig. 5.13) were 2-3°C higher than during the previous cruise. They decreased southward from 28.5°C along the northern coast to about 25°C in Delagoa Bay. Along the northern coast south to Quelimane the lowest temperatures were recorded close to the coast while usually the contrary was observed further south. The salinities (Fig. 5.14) along the northern coast south to about 18°S were slightly higher than in September. Both the salinity as well as the temperature distribution gave some indications of an eddy over St. Lazarus Bank. On Sofala Bank the salinity was approximately the same as in September, but the tongue of lowsalinity water was missing. As in September, more saline water was observed coming in from east off Inhambane. In Delagoa Bay the surface salinity was slightly higher compared to the September distribution.

SECTION I (Fig. 5.15) the salinity distribution showed an apparent ascending of the Subtropical surface water. Indications of this, however, were found neither in the temperature nor in the oxygen distribution. The maximum southward flow was probably about 130 km seaward. There was some indication of a north-going current in the inner part of the section, with a possible reversal at about 200 m.

As observed during the previous cruise, the strongest southward current occurred about 50 km off the coast in SECTION II (Fig. 5.16). More seawards, a cyclonic eddy was indicated. The northward current at the outermost stations might have been the inner part of another cyclonic eddy.

Due to bad weather conditions the outermost stations of SECTION III (Fig. 5.17) were only carried out by use of bathythermographs. The strong current shear indicated by the temperature distribution might therefore have been caused by this change in observation method.

SECTION IV (Fig. 5.18) showed a weak baroclinic structure, with indications of northerly currents of low velocities along the whole section. As also observed during the previous cruise the southward transport through this section seemed to be negligible.

Also in SECTION V (Fig. 5.19) the baroclinic structure was weak, and the southward transport seemed to be small. The distribution of surface salinity and temperature (Figs. 5.13-5.14), however, indicated transport of water from east and north-east into the section.

SECTION VI was not carried out during this cruise due to lack of time.

Figs. 5.20-5.23 show the four hydrographic sections off the Zambezi River. The effect of the freshwater outflow is clearly seen. A strong vertical haline stratification is seen at the innermost station of SECTION 3. Further out, mixing took place, resulting in only weak or non-haline stratification. The sections show the typical wedgeshape of a freshwater generated coastal current with northward transport. The influence of the freshwater outflow goes down to the bottom as far out as about 50 km from the coast.

The thickness of the mixed and homogeneous upper layer was about 25 m along most of the coast, except in the northernmost part where it reached down about 100 m. The depth of the thermocline increased gradually offshore, but in the whole area it was shallower than on the previous cruise.

January-March 1978.

During cruise no. 3 in February-March 1978 (Fig. 5.24) the surface temperatures were in general 1-3°C higher than during cruise no. 2. The highest temperatures were found at Sofala Bank where values more than 30°C occurred. In Delagoa Bay the temperature was about 27°C. Except for the southern part of Sofala Bank the lowest temperatures were found close to the coast.

A pronounced decrease in salinity compared to November was observed along the whole coast during February-March (Fig. 5.25). Except for the area between Pemba and Cabo Delgado the salinity in the zone nearest the coast was below 35 ‰. The lowest salinity, about 30‰, was observed in Beira Bay.

SECTION I (Fig. 5.26) showed the typical feature of a coastal upwelling with an intensification of the southward current close to the coast. The geostrophic velocity in 500 m relative to the surface was about 230 cm/s northward at the inner stations. It would seem to be unlikely to have had such southward velocities at the surface, which means that there was probably a reversal of the current with north-going transport in the deeper layers along the coast.

Indications of ascension of water from the deeper layers when approaching the coast were also present in SECTION II (Fig. 5.27). The core of the south-going current seemed to occur about 80-90 km off the coast. The innermost part of the section had apparently a northgoing current.

SECTION III (Fig. 5.28) probably had a northward current at the inner stations and mainly southward currents in the rest of the section. The maximum southward current seemed to be about 70 km from the coast.

SECTION IV (Fig. 5.29) shows the typical wedge-shape salinity distribution of a north-going coastal current over the shelf. An apparent lifting of the sub-surface water close to the slope was visible. The southward current seemed to have its maximum values 80-90 km from the coast.

SECTION V (Fig. 5.30) indicated a cyclonic vortice at depths below the thermocline. The maximum southward current seemed to be associated with the surface salinity minimum.

Also in SECTION VI (Fig. 5.31) the feature of a cyclonic eddy seemed to be present with north-going currents near the coast and the strongest southward currents about 100 km offshore.

The hydrographic sections off the Zambezi River are shown in Figs. 5.32-5.35. As observed on the previous cruise, the strongest haline stratification was found in SECTIONS 1 and 2. The vertical temperature gradient close to the bottom over the shelf was more pronounced than in November.

The depth of the homogeneous layer was about 25 m along the entire coast, increasing to 50 m further offshore. The thickness of this layer was less than during cruise no. 1 and rather similar to that observed during cruise no. 2 in November.

April-June 1978

In April-June 1978 the maximum surface temperature (Fig. 5.36) of about 30°C was found off the northern coast decreasing to about 22°C in Delagoa Bay. The lowest temperatures was recorded close to the coast.

Surface salinities (Fig. 5.37), above 35‰ were not observed north of Angoche. At Sofala Bank there seemed to be slightly lower salinities compared to February-March, while in Delagoa Bay they were slightly higher. As observed on previous cruises there was water of high salinity penetrating towards the coast from east off Imhambane.

As also seen during the previous cruise, the lifting of the isotherms near the coast was pronounced in SECTION I (Fig. 5.38). The vertical distribution of the geostrophic velocity at the two inner stations gave a northward velocity of about 230 cm/s in 300 m relative to the surface. It is therefore reasonable to suggest a reversal of the southward current near the coast at a depth of 150-200 m with a north-going current below this. The northbound current seemed to reach the surface just outside the wedge of lowsalinity water. Further east, southward currents seem to dominate.

Also in SECTION II (Fig. 5.39) the geostrophic velocity distribution at the two inner stations revealed a turning depth of the current with north-going currents down to about 150 m and southward currents below.

SECTION III (Fig. 5.40) appeared to have north-going currents over the shelf and slope, and southward currents further offshore.

In SECTION IV (Fig. 5.41) a core of lowsalinity water was seen 60-70 km from the coast. The same observation was made in September 1977, but that time the core was closer to the shore. If this was not an isolated patch of water, the surface salinity distribution indicate that it must have been brought into the area from the north-east. If it was a part of a tongue of low-salinity water from the area between Beira and Zambezi River it should also have been detected in the surface salinity samples. There is possibly a northward current over the shelf. The rest of the section shows southerly currents with a possible maximum in velocity about 140-150 km offshore.

Both in SECTION V (Fig. 5.42) and in SECTION VI (Fig. 5.43) there is a northerly current close to the coast.

The Zambezi sections (Figs. 5.44-5.47) show more or less the same feature as on the previous cruises, with northerly a current over the shelf.

5.6 Discussion

The seasonal variation in temperature seemed to be between 4 and 6°C along the coast of Mozambique. The highest seasonal variations were found in Delagoa Bay and the lowest off the northern part of the coast. There seemed to be a decrease in temperature from north to south throughout the year. Usually the temperature increased while moving away from the coast, except at the southern part of Sofala Bank and in the area between Pebane and Angoche in February-March. These results agree nicely with those of WYRTKI (1971).

The maximum salinities in the coastal waters of Mozambique were observed in November and the minimum in March-April. This is in correlation with the seasonal variation of the freshwater outflow (Fig. 5.1). The lowest salinities were found at Sofala Bank where a large proportion of the area had surface salinities below 30‰.

The surface salinities in the water transported by the South Equatorial Current seemed to have minimum values during March-June and maximum values from November to February (WYRTKI, 1971). In general, the surface salinities off Mozambique de creased seawards. The seasonal salinity variations seemed to be dominated by the local freshwater outflow which reached its maximum usually in February and its minimum in September-October (Fig. 5.1).

Along the northern coast south to Angoche the typical feature of a coastal upwelling appeared to occur during the north-east monsoon from November to April. As seen in Table 5.1 the conditions were favourable for a wind-induced coastal upwelling during that time. The upwelling seemed to start in SECTION I in November (Fig. 5.15) and in January (Fig. 5.26) it was fully developed. It seemed to last until April (Fig. 5.38). The upwelling was confined to the inner 30-40 km.

Several authors have shown that the characteristic length scale of a coastal upwelling might be expressed by the baroclinic radius of deformation. Off the northern coast of Mozambique this appeared to be in the order of 30-50 km and is thus in good agreement with the observed width of the upwelling zone.

From theoretical and numerical models of coastal upwelling (e.g. ALLEN, 1973, O’BRIEN and HURLBURT, 1972) as well as from observations (SMITH, 1974) it is expected that the longshore currents are mainly geostrophic. Fig. 5.48 shows the vertical geostrophic velocity profiles relative to the surface from the inner station pair of SECTION I in January and April 1978. These baroclinic components of the longshore velocity make it reasonable to believe in a reversal of the current in the deeper layers, otherwise unrealistic southward velocities would occur at the surface. The hydrographic data give some support to a “guesstimate” of about 200 m as the turning depth of the current. This would give a southward surface current of 120-140 cm/s and a counter current below 200 m of approximately the same order.

A common feature of coastal upwelling regions seems to be a baroclinic surface coastal jet and a subsurface counter current. The coastal surface jet is predicted by multi-layer numerical models (O’BRIEN and HURLBURT, 1972, McNIDER and O’BRIEN, 1973) and in the continuously stratified model of ALLEN (1973). The model of HURLBURT and THOMPSON (1972) manages to produce a realistic subsurface counter current. Observational evidence for a near-surface coastal jet and a subsurface counter current has been presented by MOOERS et al (1974) from the Oregon coast, and by JOHNSON et al (1975), from the Canary Current upwelling region.

The velocities of both the coastal jet and the subsurface counter current off the northern part of Mozambique appeared to be larger than observed in other upwelling regions. This is specially so for the undercurrent which in other areas is reported by several authors to have a current speed of 5-20 cm/s.

In April (Fig. 5.38) the vertical geostrophical velocity distribution revealed a northward current at the surface on the seaward side of the coastal jet. MOOERS et al (1976) combined direct current observations and hydrographic data to find the absolute geostrophic velocity field in the upwelling region off Oregon. The zero isotach of his section off Depoe Bay makes a pronounced upward bend on the seaward side of the coastal jet. This might indicate that the counter current can reach the surface at the seaward side of the jet, as probably is the case in SECTION I in April (Fig. 5.38). This feature is not seen in the numerical models. A reason for this might be that the different models all deal with an ocean initially at rest and thus ommitting the effects of the more permanent ocean currents.

During cruises nos. 1 and 2 in September and November respectively, there seemed to be a retroflexion of the southward Mozambique Current off the southern part of Sofala Bank. The southward transport through SECTION IV (Figs. 5.10 and 5.18) appeared to be negligible during this time. The surface salinity distributions (Figs. 5.6 and 5.14) support the concept of a recirculation at about 21 °S. Water from the east feeds the Mozambique Current between 22°S and 25°S. In February and May there seem to be a significant southward transport through SECTION IV, but negligible through SECTION V in May.

The classical concept of the Agulhas Current as an extension of the Mozambique Current has been questioned by MENACHÉ (1963) who claims that the Mozambique Current did not penetrate into the area south of 25°S at all during October-November 1957. HARRIS (1972) from his analysis of the data from spring 1964 concludes that only about half of the transport passing SECTION V consisted of water carried by the Mozambique Current from the north. The other half was water which had flowed north up the west coast of Madagascar. LUTJEHARMS (1976) treated all available hydrographic data during the north-east monsoon season. His conclusion was that during this season only part of the surface flow of the Mozambique Current runs through the Mozambique Channel, with a large volume recurving at the southern mouth to flow northward in the middle of the Channel. He also found that at s_t surface 26.8 situated of about 400-500 m, there was no continuous flow through the Mozambique Channel. The Mozambique Current below this depth would therefore derive all its water from the East Madagascar Current feeding the Mozambique Current at about 18°S.

From this it might be concluded that during most of the year a large proportion of the southward transport by the Mozambique Current does not pass the southern mouth of the Mozambique Channel, but recirculates northward further east in the Channel. This retroflexion seems to occur between 22° and 25°S and is probably weakest during the north-east monsoon season. This retrogression is followed by a feeding of the Mozambique Current by water from the east, which also appears in the investigations of HARRIS (1972) and LUTJEHARMS (1976). Further evidence of this transport is also given by GRÜNDLINGH (1977) from drift observations of a satellite-tracked buoy in August-October 1975, and he suggests an explanation in terms of a westward deflection of the East Madagascar Current.

Eddies were observed on several occasions. In the area off Angoche (SECTION II) in September (Fig. 5.8) there seemed to be a cyclonic eddy on the seaward side of the south-flowing Mozambique Current. An apparent eddy could also be seen in November (Fig. 5.16). In February (Fig. 5.27) and in April (Fig. 5.40), however, the dome shape of the isotherms might have been caused by an intensification of the northward current in the region nearest the coast, combined with local upwelling. HARRIS (1972) found a small cyclonic eddy in the same area in the spring. This eddy might be caused by a seamount slightly north of SECTION II at about N16°, E41°31’ reaching up to 230 m. It also seems reasonable to believe in local small-scale eddies over Paisley Seamount and St. Lazarus Bank further north. The grid of stations was not dense enough to locate such eddies but some indications were observed in the surface distribution of temperature and salinity over St. Lazarus Bank in November.

Eddies also seem to be a characteristic feature of SECTION V, as seen in Fig. 5.11 from September and in Fig. 5.30 from March. The cyclonic eddy seen also appeared in the data of LUTJEHARMS (1976) and GRÜNDLINGH (1977) and seems to be topographically induced. The surface salinity distribution in March (Fig. 5.25) and in May (Fig. 5.37) both indicate a cyclonic eddy in Delagoa Bay off Maputo. This eddy was also found in June 1961 by ORREN (1963).

The main features of the coastal hydrography off Mozambique can be summarized as follows:

- Along the northern coast south to about 16°S coastal up-welling occurs, resulting in a southward coastal jet and a northward counter current during the north-east monsoon season.

- A retroflexion of the southward Mozambique Current occurs between 22° and 25°S. This retrogression seems to be weakest during the north-east monsoon season and is followed by a feeding of the Mozambique Current with water from the East Madagascar Current.

5.7 Tables and figures

Table 5.1. Frequency distribution of wind direction Aug. 1977 - June 1978 (Number of observations within each sector. C = calm).

Pemba
	C	N	NE	E	SE	S	SW	W	NW
Aug
Sep
Oct
Nov	1	0	10	53	13	4	3	3	1
Dec	1	6	46	18	12	0	0	8	0
Jan	6	23	18	28	1	2	0	5	4
Feb	10	16	13	29	1	0	3	6	3
Mar	14	2	8	17	14	14	9	12	1
Apr	14	0	3	12	17	18	18	6	2
May	2	0	0	0	12	57	14	7	0
Jun	0	0	0	0	10	52	20	7	0

Lumbo
	C	N	NE	E	SE	S	SW	W	NW
Aug	16	1	2	13	6	35	20	0	0
Sep	7	2	12	35	13	12	9	0	0
Oct	5	2	16	18	8	14	4	0	0
Nov	3	9	13	35	14	12	1	1	0
Dec	21	4	20	36	7	2	1	0	1
Jan	19	10	25	20	7	2	0	2	6
Feb	29	6	16	19	4	2	0	3	1
Mar	33	10	9	15	10	4	2	6	0
Apr	34	3	1	9	10	13	12	8	0
May	30	0	0	1	6	34	17	5	0
Jun	19	0	0	0	5	26	35	4	0

Quelimane
	C	N	NE	E	SE	S	SW	W	NW
Aug	18	6	14	22	15	6	5	1	0
Sep	7	8	25	35	10	2	1	0	1
Oct	5	7	16	46	17	0	0	0	1
Nov	8	11	17	41	7	2	2	0	1
Des	8	17	26	19	9	1	3	2	5
Jan	10	10	11	17	15	17	6	2	4
Feb	14	11	13	15	12	5	9	2	2
Mar	17	10	8	18	14	9	3	9	2
Apr	21	11	4	16	11	13	10	4	0
May	31	6	2	19	15	4	9	3	3
Jun	12	3	2	4	22	20	22	5	0

Beira
	C	N	NE	E	SE	S	SW	W	NW
Aug	1	4	9	12	24	21	12	7	1
Sep	1	2	7	23	29	17	2	6	1
Oct	1	1	6	28	38	11	4	2	1
Nov	0	7	9	25	26	13	4	2	2
Des	0	3	14	24	27	13	5	6	1
Jan	3	2	10	9	23	24	3	17	1
Feb	1	2	7	9	18	23	8	10	4
Mar	0	5	6	9	21	29	16	6	1
Apr	1	4	6	9	13	24	14	14	5
May	2	8	6	13	20	17	5	9	13
Jun	2	3	4	4	8	20	21	26	2

Vilanculos
	C	N	NE	E	SE	S	SW	W	NW
Aug	20	10	2	8	10	24	13	3	3
Sep	8	9	21	29	8	13	2	0	0
Oct	7	20	31	4	18	12	0	1	0
Nov	2	25	8	22	13	18	0	0	1
Dec	5	25	9	17	13	23	1	0	0
Jan	18	11	7	16	9	32	0	0	0
Feb	9	12	8	18	6	26	3	1	1
Mar	10	10	6	15	14	33	0	5	0
Apr	24	6	3	10	6	31	3	7	0
May	39	10	7	11	3	9	3	8	2
Jun	16	2	0	10	3	42	5	12	0

Maputo
	C	N	NE	E	SE	S	SW	W	NW
Aug	10	12	13	10	6	15	14	3	9
Sep	7	11	17	25	4	13	9	1	3
Oct	5	13	23	24	6	15	3	0	4
Nov	1	15	11	29	11	20	3	0	0
Dec	6	8	31	15	9	18	4	0	2
Jan	9	4	13	12	8	31	13	0	3
Feb	9	7	14	11	6	18	15	0	4
Mar	19	7	11	24	7	10	10	1	4
Apr	9	8	14	5	3	15	15	0	10
May	12	26	7	7	6	10	13	4	7
Jun	15	8	5	10	8	6	22	11	5

Fig. 5.1. Average freshwater runoff from six localities (Fig. 2.2.).

Fig. 5.2. SECTION V - 23 May 1978.

Fig. 5.4. s_t - t relation at SECTION V - 23 May 1978

Fig. 5.3. Temperature - salinity relation at SECTION V - 23 May 1978.

Fig. 5.5. Surface temperature - September 1977.

Fig. 5.6. Surface salinity - September 1977.

Fig. 5.7. SECTION I - 1 September 1977.

Fig. 5.8. SECTION II - 8 September 1977.

Fig. 5.9. SECTION III - 18 September 1977.

Fig. 5.10. SECTION IV - 22 September 1977.

Fig. 5.11. SECTION V - 25. September 1977.

Fig. 5.12. SECTION VI - 27. September 1977.

Fig. 5.13. Surface temperature - November 1977.

Fig. 5.14. Surface salinity - November 1977.

Fig. 5.15. SECTION I - 1 November 1977.

Fig. 5.16. SECTION II - 12 November 1977.

Fig. 5.17. SECTION III - 15 November 1977.

Fig. 5.18. SECTION IV - 26 November 1977.

Fig. 5.19. SECTION V - 29. November 1977.

Fig. 5.20. Zambezi section 1-22 November 1977.

Fig. 5.21. Zambezi section 2-22 November 1977.

Fig. 5.22. Zambezi section 3-23 November 1977.

Fig. 5.23. Zambezi section 4-23 November 1977.

Fig. 5.24. Surface temperature - February - March 1978.

Fig. 5.25. Surface salinity - February - March 1978.

Fig. 5.26. SECTION I - 31 January 1978.

Fig. 5.27. SECTION II - 14 February 1978.

Fig. 5.28. SECTION III - 16 February 1978.

Fig. 5.29. SECTION IV - 23 February 1978.

Fig. 5.30. SECTION V - 7 March 1978

Fig. 5.31. SECTION VI - 13 March 1978.

Fig. 5.32. Zambezi section 1-26 January 1978.

Fig. 5.33. Zambezi section 2-26 January 1978.

Fig. 5.34. Zambezi section 3 - 23 January 1978.

Fig. 5.35. Zambezi section 4-22 January 1978.

Fig. 5.36. Surface temperature - April - June 1978.

Fig. 5.37. Surface salinity - April - June 1978.

Fig. 5.38. SECTION 1-10 April 1978.

Fig. 5.39. SECTION II - 17 April 1978.

Fig. 5.40. SECTION III - 24 April 1978.

Fig. 5.41. SECTION IV - 18 May 1978.

Fig. 5.42. SECTION V - 23 May 1978.

Fig. 5.43. SECTION VI - 25 May 1978.

Fig. 5.44. Zambezi section 1-3 May 1978.

Fig. 5.45. Zambezi section 2-2 May 1978.

Fig. 5.46. Zambezi section 3-1 May 1978.

Fig. 5.47. Zambezi section 4-1 May 1978

Fig. 5.48. Geostrophic velocity profile relative to the surface in SECTION I during January and April 1978.