Many hauls for zooplankton have been made in the Pacific and Indian Oceans. Those made in the Pacific are well summarized by Reid (1962b) and those in the eastern tropical Pacific by Blackburn (1967). The metre nets were hauled from about 200 m to the surface (ranging from 100 m to 300 m on some occasions) and the catches were expressed as ml/1000 m³ displacement volume. Observations in the Peru Current were confirmed by the large quantities of information on Hensen net hauls from 50 m to the surface (Flores, 1967; Flores and Elias, 1967 and Guillen and Flores, 1967). The mesh size of the metre nets was usually about 0.25 to 0.31 mm and the net was hauled at about 1 m/sec. So a proportion of the zooplankton population escaped through the meshes and a further proportion evaded capture. Further, it is possible that the displacement volume included an unspecified volume of algae. There is no information about the mesh selection of these nets with respect to the composition of the displacement volume. It is hoped that the loss of zooplankton through the meshes is balanced by gain of algae by clogging. Loss by escape has not been measured, although shown for some larger animals (Fleminger and Clutter, 1965; McGowan and Fraundorf, 1966), so it is possible that euphausiids everywhere are improperly sampled. However, with the present state of zooplankton sampling, the nets used probably represent the best compromise for estimates of zooplankton displacement volume over extensive areas. To count the animals and estimate their volumes from an array of nets, each sampling a band of size properly, would take so much time that the results could not have been obtained as quickly for such extensive areas.
In the Indian Ocean, Wooster, Schaefer and Robinson (1967) gives samples from a 0.200 m water column from the Indian Ocean standard net, which is like that used in the Pacific, with a mesh size of 0.33 mm and a mouth opening of 1 m² (Currie, 1963). The Indian Ocean net was hauled off Somali and southwest Arabia in strong winds and so an average wire angle of 45° has been assumed (from my own observations on R.R.S. DISCOVERY). A similar net was used by Tranter (1962) off northwest Australia and off Java. The Russians working in the Indian Ocean have used a large Juday net (0.5 m², with a mesh of 0.26 mm) and they express their results as mg/1000 m³ wet weight, essentially the same form of expression as used in the American and Australian work. Tranter (1963) has shown that the Juday net with a mesh of 0.26 mm catches more than the Indian Ocean net, with a mesh size of 0.33 mm but it is not stated whether the added living material was of nauplii or algae. Frontier (1963) has made some observations in the Guinea Current off Abidjan with a Hensen net (0.33 m²; 0.41 mm mesh). There are unfortunately no observations of zooplankton made in the Canary Current or in the Benguela Current except those made on the Meteor Expedition. On that expedition numbers of "metazoa" were counted from 4 litres of water taken at the surface, at 50 m and 100 m (Hentschel, 1933). Kamshilov's (1951) formula converting length (0.6-1.0 mm) to weights for copepodites (0.6-1.0 mm) and nauplii (0.08-0.1 mm) has been used. This procedure gives results which are comparable with those from other upwelling areas. Therefore the Pacific and Indian Ocean nets appear to sample the zooplankton populations adequately, excluding larger forms like euphausiids which were not caught by Hentschel's waterbottles. However, the two methods are so profoundly different that they are separated all through the estimating procedures.
If possible the estimates of the stock of zooplankton as ml/1000 m³ should be converted to measures of production. The first step is to determine the depth from which the nets were hauled, so that ml/1000 m³ are converted to ml/m² in a specified layer; fortunately the authors quoted above give the depths of sampling in the upwelling areas specified. An upwelling area might be 800 km in length by 200 km in width, with a current moving towards the equator at 20 km/d (Wooster and Reid, 1963). So there is an input of zooplankton from the poleward end and it is generated from the coast by upwelling. Very roughly it takes one zooplankton generation for water and zooplankton to be drifted across the length of an upwelling area. The survival of animals across the upwelling area depends on the algal production. So we may consider the generation of zooplankton to be autonomous within the upwelling area. The production of zooplankton is given by the average standing crop as ml/m², multiplied by the number of generations during the upwelling season.
Marshall and Orr (1955) give the duration of copepodite stages as a proportion of the duration of copepodite stage I. McLaren (1965) gives the duration of some copepodite stages at different temperatures, so a relation between stage duration and temperature was constructed. On the basis of data given by Marshall and Orr (1955) on maturation and hatching, the duration of a full generation could be worked out at different temperatures. They do not differ much from the estimates of Heinrich (1961). The estimated generation time has been arbitrarily lengthened by one third, to take account of the intermittent character of upwelling, due to variation in the wind stress (California, Department of Fish and Game, 1953). The zooplankton distributions off California (Thrailkill, 1956, 1957, 1959, 1961, 1963) reflect to some extent the transient structures of an upwelling region, possibly because the animals are vulnerable to food lack in the periods between upwellings. The same procedure should really be applied to the algae, but a short half of a week or so only reduced the rate of increase of algal production, whereas a week's halt in upwelling may cause the failure of a local brood of nauplii, and when upwelling returns it would take half a generation for a new brood to get underway. The estimate of one third of the generation time was taken from the seasonal picture of intermittent upwellings off Southern California (California, Department of Fish and Game, 1953).
It is possible that the intermittent halts in the upwelling process actually sustain greater levels of production. The rate of increase of algal production is reduced and the zooplankton production is perhaps destroyed. Where the upwelling is resumed, the algae continue to produce, but the grazing restraint is no longer there. So a new outburst is possible until the zooplankton has been regenerated. The available data in different areas are not good enough to support or deny this speculation.
In Table I, where the data are presented in detail, column L gives the generation time with one third added. In all the upwellings examined, temperature observations at the surface are available from the sources given in section 6.6. Only in a few cases can the observations be averaged adequately through a season, but they are well enough established to estimate generation time. Columns N1, and N2 give the total secondary production as tons C.106/yr, with the lengthened generation time (N1) and with an uncorrected generation time (N2), so the quantity in column N2 is one quarter greater than that in N1. The quantity ml/m² (or g/m², wet weight) is converted to carbon by the factor 1/17.85 (Cushing, 1958). This figure, raised to the area of the upwelling area (as defined above) and raised by the number of generations (as defined above), gives the quantity produced, in tons C/yr.
The secondary production in tons C.106/yr is given in Table I. Although there is a large quantity of information on secondary production in the California Current, observations in other upwelling areas are sufficient only to establish an average, but not a seasonal trend, as for the radiocarbon measurements. The most important upwelling areas in terms of secondary production are those in the Peru Current and the Benguela Current, where production is of the order of 3-5.106 tons C/yr (using column N1). The production in other upwelling areas for which there is enough information is around 1-2.106 tons C/yr. There is not enough information to draw any conclusion for the Canary Current in secondary production and that for the Guinea Current is based on very few observations. For many areas for which radiocarbon observations are available, there are not adequate observations in zooplankton. However, the southwest Arabian upwelling, the Domes off Costa Rica and Java and the Indonesian area appear to be areas of medium zooplankton production. So far as it goes this arrangement of high production in the Peru and Benguela Currents and medium production in other areas does correspond to the order given for the radiocarbon measurements.
Summarizing, the estimates of secondary production are based on the use of one standard net, or its analogues. In fact a number of different designs have been used (see Reid's table 2) but the differences are not great enough to generate differences in quantities caught. Indeed the very meticulous methods used by Hentschel on the Meteor Expedition tend to confirm the results from the CALCOFI/POFI/IIOE nets. Perhaps the assumptions that the loss of nauplii is balanced by gain of algae, and that the proportions escaping are small, are roughly justified.
