3. THE LAKE ENVIRONMENT

3.1 Environmental Measurements

3.1.1 Meteorological and Hydrological Data

Supplementary environmental data were obtained from the Provincial Hydraulic General Station in Xining. Records of lake levels, air and water temperatures and precipitation and evaporation were obtained for lakeside stations near the Heima River (36°43'N, 99°47'E) and at Yao Wuyao (36°35'N, 100°29'E) and Sha Tueshi (37°13'N, 99°51'E). Records of monthly flows (measured by submersible vanes) were obtained for the Buha River and the Shaliu River. Flows in the Buha are metered in the channel at the road bridge (37°02'N, 99°44'E) and do not include those of other distributary channels. Flows in the Shaliu are metered at Gangca (37°19'N, 100°08'E), upstream of a diversion weir.

As wind-driven surface oscillations (seiches) often cause the water at any point on the shore to rise and fall over a period of several hours, lake-level measurements must be made several times daily. In this report monthly averages for each of the three lakeside stations are used. Standard meteorological installations are used to record air temperature, precipitation and evaporation, except that a non-standard 20-cm diameter evaporation pan is used to minimise problems with ice. Water temperatures are measured morning and evening at 20–30 cm depth near the shore (and therefore may be higher than in open-water areas). Again, data herein are monthly averages for the three stations.

3.1.2 Sampling Stations

Samples in 1990–91 were obtained from seven stations (I–VII) across the lake, and in 1992 from four or five stations (1–5) near the Fish Factory, where most of the trawler catch is taken. Stations I–VII were I: Chuancangwai (midway between the factory and the northern shore), II: Erlangjian (“neck” of the lake, defining “South Bay”), HI: Haxingshan (south east of Sea-Centre Hill), IV: Satuesi (south of Sea-Centre Hill), V: Tiebuqi (north west of Three Strange Stones), VI: Sangueisi (the bay near the Buha mouth) and VII: the Buha mouth. Stations 1–5 were in the “South Bay”, near the Fish Factory, along an arc leading anticlockwise from the factory jetty. In 1992 sampling required about 3 h travelling time in the 125-hp research boat Standard records included samples for chemical analysis, benthos, phytoplankton and zooplankton. Profiles were recorded of water quality, and Secchi transparency was monitored using a 20-cm white disk.

3.1.3 Water Chemistry

Lake-water samples were obtained using a Ruttner sampler. Chemical parameters were measured on site, or soon after return to the laboratory, using a Palintest PT260 Chemical Analysis Laboratory. Measurements of conductivity, dissolved oxygen, temperature and pH were made with a WPA Environmental Multiprobe. Light penetration (400–700 nm) was measured using a Partech lake profiler.

3.1.4 Benthos

Lake sediment samples were obtained using a Peterson Grab (area 0.0625 m²) or an Orange-Peel Grab (0.05 m²). Only one grab sample was taken at each station. The samples were washed and sieved (135 μm) immediately after collection, and animals and larger algae (e.g. Cladophora) were removed and transferred to vials containing 4% formalin with a few added drops of iodine. In the laboratory the total wet-weight biomass of recovered material was measured using a photobalance accurate to 0.0001 g. The average biomass per square metre of lake sediment was estimated by multiplying the sample biomass by an appropriate conversion factor, determined by the grab sample area.

3.1.5 Plankton

Phytoplankton was sampled by lowering a 55-μm mesh, 40-cm diameter Apstein net to 5, 10 and 15-m depth, and hauling to the surface (so providing only a general indication of the depth distribution of organisms). Samples were transferred to 4% formalin with a few added drops of iodine, and returned to the laboratory. For counting, two 0.1-ml subsamples were transferred to separate microscope slides. The numbers of individuals of each genus were determined by counting 50–100 fields of vision for each slide, and a conversion table was used to estimate biomass. Estimates of average biomass per litre of lake water were calculated by estimating the volume of water in each haul.

Samples of zooplankton were obtained using a 335-μm Apstein net (40-cm diameter) at three depths, as for phytoplankton. Total numbers were counted for each sample, with individuals grouped as broad taxa (Protozoa, Rotifera, Cladocera, Copepoda and nauplii). Samples of 20–40 individuals were randomly selected for measurements of length, and estimations of cladoceran and copepod biomass followed Downing & Rigler (1984). Biomasses of other zooplankters were estimated from measurements of length and simple geometric approximations, assuming a specific gravity of unity. Conversions of sample counts to average biomass per litre of lake water were made by estimating the volume of water sampled by the respective hauls, as for phytoplankton.

3.2 Basin Morphometry

A bathymetric survey of Qinghai Lake was made in 1961–63, in the course of the survey by Academia Sinica (1979), but the data need to be adjusted because the lake level has changed. Estimates of the lake surface area for a given water level were derived by enlarging the contour map (Academia Sinica 1979: p. 8), determining the area enclosed by the 5-m isobath using a planimeter and estimating the required area by linear interpolation. Planimetry was also used to plot depth profiles of area (A_n, A_m) and volume (V_m__n). The volumes of strata (n, m) were estimated from:

The following equations, derived from the planimetric data, may be used to estimate the lake surface area (km²) and volume (billion m³) for elevations between 3191–3196 m:

Morphometric data for the lake in 1990 are included in Table 3.1, with data for 1961–63. A bathymetric map is shown in Figure 3.1, and depth profiles of area and cumulative volume are shown in Figure 3.2.

The comparative data show that the lake level fell by 1.85 m between 1961 and 1990 (see section 3.3), with corresponding reductions in surface area and volume. There has also been an apparent increase of 0.6 m in the mean depth (the ratio of volume to surface area). Note that the mean depth of 19.15 m reported by Academia Sinica (1979) cannot be correct if (as assumed here) the volume and surface area are both accurate. If the volume or area data are wrong, the error(s) will be perpetuated in the estimates for 1990.

Assuming that the maximum depth of 28.7 m reported for the lake in 1961–63 is accurate, the maximum depth (m) for any given surface elevation (m) may be calculated from:

Maximum Depth = Surface Elevation - 3167.3

Table 3.1
Morphometric data for Qinghai Lake, derived from the 1961–63 survey (Academia Sinica 1979) and the present investigation (1990).

* Averages in successive years were 3196.08, 3195.91 and 3195.76, respectively

† Calculated as volume/area (cf. 19.15 m in original report)

3.3 Historical Changes in Water Level and Salinity

The level of Qinghai Lake has been regressing for several thousand years, following a regional trend towards increased aridity, although the process has been interrupted for long periods. The first recorded measurement of the lake level is 3206.6 m in 1884–86 (Chen 1991), implying a maximum depth of 39.30 m. Other early records are 3206.5 m in 1893–94 and 3205.0 m in 1908 and again in 1927 (Dai 1987, Chen 1991). The data for 1990 indicate a fall of 12.37 m in about 105 years, at an average rate of 11.7 cm per year.

Regular measurements of the lake level commenced in 1955, with establishment of an hydrological station operated by the Qinghai Water Conservancy Bureau. Records for 1956–90 are summarised in Figure 3.3. A continuing decline is obvious, despite minor recoveries following unusually wet years, as in 1989. Over the 35-year period of record (1956–90) the average rate of decline has been 8 cm per year. From 1961–90 the level fell by 1.85 m, at an average rate of 6.4 cm per year.

It is of interest to know whether the lake's regression may have been accelerated by diversions of water for irrigation. A rough assessment is possible from data for the Shaliu River at Gangca, where in 1958–64 a 1–2 m weir and 10-m wide, 26-km long diversion canal were constructed to supply an irrigation area of 6667 ha (Qinghai Water Conservancy Bureau 1984). Water is diverted from the river throughout the ice-free period (early spring to late autumn). In 1981 the average flow in the canal was 8.27 cumecs (m³s^-1) and the total seasonal diversion was 123.5 M m³. The metered discharge of the Shaliu in 1981, including the diversion, was just over 281 M m³, so that the diversion accounted for about 30% of the river's total discharge. As the design capacity of the canal is 12 cumecs, it could potentially divert 187 M m³ over an irrigation season of 180 days. Assuming a total annual inflow to the lake of 2187 M m³ (Chen 1991), the Shaliu diversion accounted for about 5.3% of the flow in 1981, and would divert about 7.9% if operated at full capacity. On this basis, it appears that diversions have made a small but significant contribution to the fall in lake level.

Flows in the Haergai and Qianji rivers also are subject to diversions. The effects on the lake water budget may be significant for the Haergai, as its discharge rivals that of the Shaliu (section 3.4.2). Otherwise, diversions are responsible for substantial flow depletions in each of these three rivers, and apparently no efforts are made to maintain flows sufficient for local spawning by naked carp.

The lake's regression has caused the inflowing rivers to downcut their channels, adjusting their base levels to that of the lake. This is most noticeable near the road crossing on the Buha River, where stone walls have been constructed to control bank erosion. The mean annual sediment load in the Buha is 484,700 tonnes (range 23,100 to 862,000 tonnes; data from Provincial General Hydraulic Station, Xining). Erosion will have increased the sediment load in the river, supplemented perhaps by the effects on catchment soils of the herds of yaks, sheep and goats. Sediment deposited in the channel probably has affected naked carp spawning areas (section 5), and alluvium deposited in the lake has contributed to siltation around nearby Bird Island, which first appeared as a distinct peninsula on a map drawn in 1972. The delta of the Buha has grown considerably in the 20 years since the isthmus to Bird Island was formed.

The salinity of Qinghai Lake has varied over time, but historical data are few and do not indicate a general rise in salinity, as might be supposed from the fall in water level. According to Ying (1987) the salinity was 5–6 g L^-1 in some parts of the lake during the early 1950s, but this probably reflects the local effects of freshwater inflows. Recent records show salinities of 12.49 g L^-1 in 1961–63 (Academia Sinica 1979), 13.28 g L^-1 in 1982 (Provincial General Hydraulic Station data) and 14.152 g L^-1 in 1986 (Chen 1991). The present lower salinity (1990) of 12.5 g L^-1 may reflect the increase in lake level following above-average inflows in 1989 (e.g. Fig. 3.2). If the 1990 record is excluded, there does appear to be a trend towards increased salinity. This is not supported, however, by a surprisingly high salinity of 13.8 g L^-1 recorded by Schmidt in 1880 (Chen 1991). If Schmidt's analysis is representative of the lake 100 years ago, there is little reason to speculate about trends in salinity.

At any one time there is little horizontal variation in the salinity of open-water areas (Academia Sinica 1979), but intensive sampling in 1985–87 shows that samples from near-shore regions may be about 10% less saline than those from the main body of the lake (Professor Yu Shengsong, Institute of Salt Lake Studies, personal communication). Freshwater seepages occur along the lake shore near the Fish Factory, and the presence of extensive marshland areas between the mouths of the Shaliu and Haergai rivers suggests that seepage may occur also in that region. The Academia Sinica report mentions areas of groundwater inflow near Erlangjian Sand Beach and Sand Island.

3.4 Hydrology

3.4.1 Precipitation and Evaporation

A graphical summary of precipitation and evaporation data for 1970–90 is shown in Figure 3.4. Over 1970–90 the average annual precipitation (rain and snowfall) for the three lakeside stations was 377 mm (SD 64 mm), with a minimal 270 mm in 1973 and a maximal 561 mm in 1989. The average annual evaporation was 1573 mm (SD 131 mm), with extremes of 1295 mm in 1989 and 1774 mm in 1979. Over the 21 years of record, therefore, the rate of evaporation is about 4.2 times that of precipitation, a little more than the figure of 3.8 reported by Academia Sinica (1979). Over 80% of average seasonal precipitation occurs as rainfall between June and September (Academia Sinica 1979), so that summer is the season of peak river inflows. An analysis of spatial variations between the three stations is not appropriate here, but evaporation and precipitation on the southern shore tend to be higher than on the northern shore. Academia Sinica (1979) recorded annual evaporation and precipitation averages of 395 and 1487 mm respectively on the southern shore, compared with 377 and 1433 mm on the northern shore.

3.4.2 Major Inflows

In 1961–63 seven rivers contributed about 95% of the total surface inflow to Qinghai Lake (Academia Sinica 1979). Clockwise from the western shore, these are the Buha, Qianji, Shaliu, Haergai, Ganzhi, Daotang and Heima rivers (Fig. 1.1). The contribution of the Ganzhi (Ganzihe) may now be discounted, as the river became separated from the lake during the 1960s and is now sealed by a sand bar. A population of naked carp isolated in the Ganzhi has subsequently been described as a distinct subspecies (section 5).

The Buha River (Buhahe, Pu-k'o Ho) is 286 km in length, with a catchment of 16,570 km² (14,337 km² according to one report) and an estimated annual runoff of 1064 M m³ (1961–63). The river forms a braided delta about 20–30 km from the lake, and its lower reaches meander along the east-west axis of the former lake bed. It enters the lake through at least two primary channels, only one of which is metered. The metered channel, passing under the road bridge at the Buha settlement, appears to account for at least three quarters of the total discharge. Apart from the effects of bank erosion, and perhaps local grazing, there are no major urban settlements or industrial activities that might affect flows in the river. The bed is a mixture of gravel shoals and silted, muddy areas whose nature changes with the effects of spates. The nature of the substratum is important for reproduction and recruitment of young fish (section 4).

The Shaliu River (Shaliuhe) is 106 km in length, with catchment 1442 km² and annual discharge 246 M m³. It enters the lake through a delta mid-way along the northern shore. The catchment includes hilly land where there is some minor cultivation. About 15 km before the lake, the river passes through the town of Gangca (Kang-cha), a growing population centre and the seat of regional government The local irrigation area, supplied by a diversion weir, includes a large part (6667 ha) of the river delta.

The Haergai River (Haligenhe), on the north-eastern shore, is 110 km long with a catchment of 1425 km² and average annual discharge 242 M m³. Flows are limited by an irrigation scheme established in 1965, supplied from a diversion weir near the road crossing, about 20 km from the lake.

The Heima (Heimahe), Qianji (Wuha-Alanhe) and Daotang (Daotanghe) rivers are small, permanent streams with catchments smaller than 500 km², and contribute only small inflows to the lake. The Heima River passes through Heima Village before flowing through pasture land to meet the southern lake shore. Water quality in the river may be affected by sediment from the pastures and waste from cattle slaughtered in autumn. The Qianji River, on the northern shore, is impounded by a 1–2 m weir that feeds a diversion canal The weir is frequently demolished by floods, and has been rebuilt several times.

The Daotang River drains the lowlands of the ancient lake bed at the eastern end of Qinghai Lake and flows into Erhai (“Ear Lake”), a shallow freshwater lagoon separated from the lake by a sand bar. The river does not have a well-developed delta and lacks the “flashy” pattern of flows typical of other streams entering the lake. Flows in the Daotang are not metered, but may contribute as much as a quarter of the flow in the main branch of the Buha River. Water from the Daotang enters the lake by seepage through the lagoon bar rather than by surface flow. A population of naked carp apparently persisted in Erhai for some time but has been eliminated, perhaps by over-fishing or as a side-effect of chemicals used to dip sheep.

Figure 3.5 compares the annual discharges of the Buha and Shaliu rivers in 1970–90 (see Appendix II for original data). The average annual discharge of the metered channel of the Buha River was 763.62 M cumecs (se 371.30), ranging between 196.28 (1973) and 1920.13 M cumecs (1989). In comparison, the Shaliu River had an average annual discharge of 242.42 M cumecs (se 94.21) and ranged between 104.08 (1979) and 470.37 (1989) M cumecs. On average, the discharge of the Buha is a little more than three times that of the Shaliu, and contributes about half of the total surface inflow to the lake.

Table 3.2
Approximate water budget for Qinghai Lake, 1970–90. Data are mm unless otherwise indicated.

Although stream runoff in arid areas often shows wide spatial variation, flow trends in the two rivers tend to be broadly similar (annual discharges are correlated: r = 0.80). Figure 3.S shows that exceptional floods occurred in the Buha in 1971 (peak 200 cumecs in September), 1983 (186 cumecs in July) and 1989 (284 cumecs in July). Minimal flows occurred in 1973 (peak 18.20 cumecs in July) and 1978 (40.10 cumecs in July). The Shaliu also showed a 21-year maximum in 1989 (peak 58.00 cumecs in July) and a marked low in 1973 (peak 8.02 cumecs in October). These variations are clearly related to annual variations in rainfall (section 3.4.1). As the Buha, and to a lesser extent the Shaliu, provide major spawning grounds for naked carp, it would be of interest to know whether variations in river flow are correlated with spawning success, recruitment and fishery catches (cf. sections 4–5).

Figure 3.6 compares monthly flows in the two rivers over the 21 years of record. In each case there is an obvious seasonal pattern, with flows consistently lowest during the period of ice cover (November to April) and reaching a variable peak in July. The seasonal increase of flow begins in April in the Shaliu, about a month before the Buha, but the Buha rises more quickly to its peak.

3.4.3 Lake Water Budget

Table 3.2 shows an approximate water budget for Qinghai Lake, calculated from annual data for evaporation, precipitation and river discharge. In the absence of detailed information, the budget makes a number of assumptions: for example, the lake surface area is taken as 4635 km² and the combined discharge of the Qianji and Heima rivers is estimated as 46.8% of the combined discharge of the Buha and Shaliu, judged from the relative sizes of the catchment areas. The flows of smaller streams are ignored.

A comparison of the inputs and outputs indicates that the lake incurs an annual average deficit of approximately 875 mm per year. To account for the observed decline in lake level of about 64 mm per year (cf. section 3.3), an unmetered annual inflow of about 800 mm must be assumed. This could be accounted for by minor inflows and by subterranean seepage, which is likely to be significant in this region of soft permeable soils (Academia Sinica 1979 estimated annual groundwater inflows as 640 M m³, equivalent to about 140 mm per year). Seepage from the Daotang River, via Erhai, is likely to be significant and, as mentioned, other areas of freshwater seepage may exist along the lake shore.

Figure 3.7 demonstrates a clear correspondence between changes in lake level in relation to the water balance calculated in Table 3.2. By eye, a deficit of-800 mm (indicated by a broken line) appears to be the break-even point, above which the lake level increases.

Recoveries are evident in 1975 and 1983, and particularly 1989, when a major flood in the Buha produced a positive annual balance for the only time in the 21 years of record.

3.5 Lake Thermal Characteristics

3.5.1 Wind Speed and Direction

Appendix II includes details of wind speed and direction at Jianxigou (southern shore, midway between the Fish Factory and Heima Village), Gangca (northern shore) and Haiyian (north-eastern area) in 1971–90. Jianxigou and Gangca best approximate the exposed conditions across the surface of the lake, whereas conditions at Haiyian are influenced by nearby mountains. Over the course of a year, Jianxigou is exposed mainly to changeable winds from the south or south south west, without any indication of a clear seasonal pattern. Gangca is subject to winds mainly from the north west in winter and from the south east in other seasons. The prevailing wind at Haiyian is consistently from the north north west Average monthly wind speeds at all three stations usually are in the range 3–4 m s^-1, and tend to be lowest at Haiyian. The Academia Sinica (1979) survey recorded maximal monthly wind speeds of 13–22 m s^-1 in 1960–78. The fetch of the lake (its length along the axis of the prevailing wind) is about 60 km, corresponding roughly with its north-south axis. The long fetch would promote surface oscillations (seiches), given variable winds, and would also discourage development of a shallow thermocline it the main body of the lake. The prevailing wind direction is more-or-less perpendicular to the main vector of surface currents in the lake, driven by inflows from the Buha River (section 3.5.3).

3.5.2 Air and Water Temperatures

In 1970–90 average monthly air temperatures in the lake region varied from-16.1°C in January to 11.0°C in July, with extremes of-21.9°C (January 1986) and 13.5°C (July 1971). Air temperatures tended to be below 0°C from mid-October to mid-April. Average monthly surface water temperatures ranged from-0.6°C in January to 13.5°C in August, with extremes of-0.9°C (February 1980) and 15.4°C (August 1989). Water temperatures tend to be at or below 0°C from November to March. Surface ice occurs for 100–129 days each year, and typically is 0.5 m thick on the lake and up to 1.0 m on the rivers (Academia Sinica 1979). The correspondence of trends in average air and water temperatures is shown in Figure 3.8, and supporting data are shown in Appendix II.

3.5.3 Seasonal Thermal Behaviour

The Academia Sinica (1979) survey revealed that Qinghai Lake has a complex pattern of seasonal thermal behaviour, strongly influenced by currents from the Buha River. The main vector of the Buha current flows to the north of Bird Island and east to the region of Sand Island, where it turns southward, dividing into two parts: one flows south west then north west, completing a clockwise circuit, and the other flows towards South Bay and returns as a secondary back current Other back-currents are associated with inflows from streams along the northern shore, and with shallow-water areas around the periphery of the lake. The Academia Sinica team defined four horizontal circulation zones:

1. Littoral Zone (to 8–10 m depth; mixed by wave action; no stratification)

2. Shallow-Water Zone (to 8–14 m; mixed by return currents associated with wave action and the Buha current; no persistent stratification)

3. Mid-Water Zone (to 14–23 m; mixed by the Buha current; no persistent stratification)

4. Deep-Water Zone (to >23 m; stratifies seasonally, although with localised variations dependent on the influence of the Buha current)

The area of the lake basin where seasonal stratification occurs is well off-shore, and extends through the main body of the lake and into South Bay, where sampling was concentrated during this investigation.

Vertical profiles of temperature and water quality in South Bay were recorded on three occasions in 1990 and on two occasions in 1992 (no records were obtained in 1991). Sampling was not possible during the period of ice-cover (November to April), and conditions in May and October sometimes were hazardous. Data for 1990 are used here to illustrate the conditions that develop in the South Bay area. There was little variation between stations, so that data for station 2 only need be presented. Station 2 is approximately 20-min travelling time on a bearing of 315° from the factory jetty.

Temperature profiles in June, July and September 1990 are shown in Figure 3.9. In mid-June there was a weak thermocline at about 8 m, but is likely that this did not persist From April–-May until June the water column probably was mixed throughout, with the first signs of persistent stratification not apparent until early July. By 18 July there was a broad metalimnion, with a thermocline at about 18 m depth. Peak stratification would have developed in August, and by 12 September there were early signs of epilimnetic cooling, with the primary thermocline descended to about 19 m and a weak secondary thermocline at about 1 m depth. Holomixis is likely to have occurred in October.

Figure 3.10 provides supplementary data for 12 September. The salinity was uniformly 12.4–12.5 g L^-1 at all depths and pH was uniformly 9.2–9.3. However, there was significant hypolimnetic oxygen depletion (to about 50% of saturation), and light was strongly attenuated immediately below the surface. These profiles indicate a concentration of plankton in the epilimnion, and deep-water oxygen depletion caused by decomposition of organic matter.

Winter profiles were not recorded in this investigation, but the Academia Sinica (1979: Fig. 27) data for January 1962 shows distinct inverse stratification, with temperatures near zero at the surface, a thermocline at 5 m and temperatures of 2–5 °C at greater depths. It is clear that Qinghai Lake has two periods of stratification and two periods of holimixis in each annual cycle, and therefore is a dimictic lake.

3.6 Chemistry

3.6.1 Major Ions, Salinity and pH

According to Academia Sinica (1979), the water of Qinghai Lake is chemically dominated by sodium chloride (66.36%), with secondary amounts of magnesium sulphate (19.80%). The 1961–63 survey indicated no significant horizontal variations in chemical composition, although there was some minor seasonal variation associated with the formation and melting of ice and photosynthetic depletion of bicarbonate concentrations over summer. The survey also recorded pH values in the range 9.1–9.4, indicating high alkalinity.

The few chemical data gathered during the present survey were in general agreement with those reported by Academia Sinica. The lake salinity throughout 1990–92 varied little from 12.5 g L^-1, so that the chemical analyses made in 1961–63 are likely to be a valid indication of present conditions. No significant horizontal or vertical variations in salinity were recorded from the four offshore stations in South Bay, and values of pH also were consistently near 9.2. Although the salinity increased to about 14.2 g L^-1 during the mid-1980s (Chen 1991), unusually wet conditions in 1989 increased the water level (section 3.4) and reduced the salinity. Table 3.3 compares data for 1962 and 1986.

The general excess of evaporation over precipitation, runoff and seepage indicates that the lake will tend to increase in salinity over time. The Academia Sinica (1979) team estimated that evaporation tends to increase the salinity by 0.73 g L^-1 annually, and that inflows reduce salinity by 0.22 g L^-1, suggesting a net annual increase of about 0.5 g L^-1. Any trend is masked, however, by variations in climate, and long-term data are required for proper evaluation. The presumed annual increase is consistent with salinity variations over the past three decades, but not with Schmidt's reported salinity of 13.8 g L^-1 in 1880 (section 3.3). If this report is wrong, the recent data suggest that the lake would have been quite fresh in 1880, and that it will increase its salinity twofold in another 25 years.

3.6.2 Nutrients

The nutrients in lake water determine its potential fertility, or trophic status. The distinction between potential and actual fertility is useful because the latter may be constrained by factors not related to nutrient availability (for example, low temperature or high abiogenic turbidity). Phosphorus and nitrogen are the nutrients most likely to be in limiting supply, but minor elements like silicon and iron may also be implicated. Bioassay experiments are required for confirmation.

Nutrient determinations recorded in the present survey were essentially ‘spot checks’ to confirm that levels did not vary greatly from those recorded in 1961–63. The Academia Sinica survey reported an average inorganic nitrogen concentration of 0.08 mg L^-1, including 0.04 mg L^-1 as ammonium, 0.036 mg L^-1 as nitrate and 0.004 mg L^-1 as nitrite. Nitrogen levels were highest in spring (0.162 mg L^-1, mainly as ammonium) and least in summer (0.030 mg L^-1), undoubtedly reflecting utilization by phytoplankton. The average concentration of dissolved phosphorus compounds was 0.02 mg L^-1, and seasonal trends were similar to those for nitrogen, showing depletion during the growing season. Levels of silica (SiO₂) averaged 0.35 mg L^-1, and iron averaged 0.016 mg L^-1. The report notes that the nutrient concentrations of interstitial water from the lake sediments were 15–30 times higher than the lake water, suggesting that the rate of mud-water nutrient exchange, which must be limited by low temperatures, is an important constraint for lake-water concentrations. The Academia Sinica team concluded that the lake is mesotrophic, but this judgement was influenced by assessments of plankton biomass. In terms of nutrient levels alone, Qinghai Lake should be regarded as oligotrophic.

Table 3.3
Chemical composition of Qinghai Lake water (g L^-1).

¹ Academia Sinica (1979)
² Sino-Australian-Swiss Investigation, cited by Chen (1991)

Table 3.4
Nutrient analyses for Qinghai Lake, 10 September 1992. Stations 1–4 are in South Bay. Station “5” was a “near-shore” station about 500 m north-west of the factory jetty and 500 m offshore. Not detectable: nd

Too few nutrient assays were made to compile a clear seasonal picture. Determinations in September 1992 shown in Table 3.4, suggest higher concentrations of phosphate and ammonium-nitrogen than those reported by Academia Sinica (1979), and the variation in nitrate-nitrogen between stations also is high. Spot measurements of water from the South Bay on 27 April and 28 July 1991 showed 0.07 and 0.71 mg L^-1 ammonium, 0.043 and 0.063 mg L^-1 nitrate and 0.06 and 0.10 g L^-1 phosphate, respectively.

3.7 Lake Sediments

The superficial sediments of Qinghai Lake include mixtures of sand, silt and clay. Some organic matter is present, although measurements have not been made, and sediment-grab samples from deep water (>20 m) in the South Bay are varved (layered), indicating little disturbance by currents. Other information about the chemistry of the lake sediments is given by Academia Sinica (1979), and Chen (1991) provides a general introduction to the literature on paleoclimates inferred from sediment cores. The Academia Sinica survey showed that the region of the mud-water interface generally is oxygenated, but that conditions below 2–3 cm depth are reducing. Values of redox potential (E_h) generally decrease from the lake margins towards the centre, and areas where the sediment E_h is <100 mV cover about 40% of the total lake area. The area of lowest E_h (<200 mV) is in South Bay. In general, the composition and electrochemistry of the superficial sediments are influenced by currents produced by river inflows, particularly the Buha River.

3.8 River Water Quality

Throughout 1990–92, occasional spot determinations were made of water quality in the principal inflowing streams, with regard for salinity, temperature, oxygen content and nutrient concentrations. This information was used mainly to support observations regarding spawning (section 4). A sample of river water quality data obtained in April 1992 is shown in Table 3.S. Nutrient levels generally showed wide variations, but it is not possible to determine meaningful trends from so few data. Nitrate concentrations in the Buha, Heima and Shaliu rivers tended to be about an order of magnitude higher than those in the lake, due perhaps to runoff from pasture and human settlements. In general, there were no signs of contamination sufficient to cause concern.

Table 3.5.
River water quality measurements, 15 April 1992. Data as mg L^-1 unless otherwise indicated.

3.9 Benthos

The benthos, and specifically the zoobenthos, of Qinghai Lake is of special significance because it is a primary source of food for adult naked carp. About 80% of the biomass is represented by chironomids (Insecta: midge larvae), overwhelmingly of the Tendipes reductus complex, and amphipods (Crustacea: scuds) also are conspicuous. The species of Gammarus is tentatively identified as Gammarus nr lasahensis (Mr J.H. Bradbury, University of Adelaide, personal communication). The Academia Sinica survey in 1961–63 recorded 19 taxa, including Tendipes and Gammarus. Otherwise, insect taxa included the chironomids Cricotopus silvestris (complex), Cryptochironomus, Heptagia, Psectrocladius and Psilotanypus (Diptera: Chironomidae), the brine fly Ephydra (Diptera: Ephydridae), the beetles Agabus (?) and Laccophilus (Coleoptera: Dytiscidae) and the water boatman Corixa (Hemiptera: Corixidae). Gastropod molluscs (snails) included Choanomphalus, Glaba pervia, Gyraulus grodleri, Radix logotis and R. ovata. Oligochaete worms included Limnodrilus helvetica and species of Nais and Paranais. In addition, ostracods (Crustacea) and nematodes (Nematoda), both usually benthic forms, were common in plankton samples (large numbers of ostracod shells were also noted in sediment samples). The Academia Sinica team regarded this as a conservative list, as few collections were taken from littoral areas.

In 1990–92 grab-sample collections from South Bay contained only chironomids and amphipods. The absence of other species may reflect the methods used and the paucity of samples, so that it is not possible to draw conclusions about changes in the fauna. Nevertheless, it is remarkable that no gastropods or ostracods were observed in samples, either from deep=water or shallow marginal areas. In samples taken from near-shore areas in September 1992, again only chironomids and amphipods were found. However, three pulmonate gastropods, including a species of Gyraulus, were found in freshwater seepages on the lake shore immediately west of the Fish Factory. The Academia Sinica report also lists pulmonates (“lung-snails”) from this region of the lake. Estimates of benthic biomass at various stations in 1990–92 are shown in Table 3.6. The Academia Sinica (1979) investigation recorded a comparable mean biomass of 400 organisms m^-2 or 973 mg m^-2 (wet weight). In 1991 no useful data were obtained for benthos. The available estimates are consistent with the lake's apparent oligotrophic status, but the high variability suggests that the benthos is not distributed uniformly, perhaps reflecting variations in grain size, redox potential or other sediment characteristics. The naked carp probably exploit localised or patchily-distributed areas that are more productive than these data suggest.

Table 3.6
Benthic biomass (mg wet weight m^-2) at lake stations in 1990–92, estimated from single grab samples. The asterisk indicates that only a single Gammarus was recovered.

I: Chuancangwai — midway between factory and northern shore; II: Erlangjian — neck of lake; HI: Haxingshan — south east of Lake Centre Hill; IV: Satuesi — south of Lake Centre Hill; V: Tiebuqi — north west of Three Strange Stones; VI: Sangueisi—bay near Buha mouth; VII: Buha mouth.

According to Academia Sinica (1979), near-shore areas with muddy bottoms are more productive of benthos than deep-water, sandy areas, and the south-western region (near the Buha mouth) is more productive than the north-eastern region. Deep-water areas, like those sampled in the present survey, were comparatively sterile. Chironomids were conspicuously more abundant near the mouths of the rivers, and densities of >1000 larvae m^-2 were recorded at a number of stations.

The 1990–92 grab data may not be an accurate indication of benthic biomass or species composition, but do suggest that deep-water areas are sparsely populated. The main feeding grounds for naked carp are likely to be off the mouths of the rivers. It is possible that mass emergences of chironomids in summer may severely deplete the available food supplies and encourage the fish to diversify their diet.

3.10 Plankton

Phytoplankton and zooplankton data are essentially spot measurements, as for benthos. Species composition was not monitored, but biomass (mg L^-1) was measured for samples taken in July, August and September 1990 and May and September 1992 (Table 3.7). There are no data for 1991.

The records in Table 3.7 are consistent with generally low productivity, implied by the oligotrophic character of the lake. They also suggest that the biomass of plankton in near-surface tows is comparatively low, suggesting avoidance of high light intensity and disturbance by waves. Although data for the rivers are not included here, it is curious that the biomass of phytoplankton in the inflowing rivers tends to be an order of magnitude greater than in the lake (cf. also Academia Sinica 1979), as there are no apparent slow-flowing or lentic areas that would encourage phytoplankton populations. The converse is true—as would be expected—for the larger zooplankton species (>350 mm).

The Academia Sinica (1979) survey recorded 35 genera of phytoplankton (Appendix HI), among which the diatoms Cyclotella, Navicula, Nitzschia, Amphora and Cocconeis, the chlorophytes Cladophora and Oocystis and the euglenophyte Trachelomonas occurred year-round. According to the survey data, horizontal variations in the distribution of phytoplankton (and zooplankton) are most conspicuous in summer, and are related to substratum and depth. More diverse assemblages occur in mud-bottom areas than in sand-bottom areas, and denser populations occur in water above 14 m, corresponding to the depth of the thermocline in summer.

Table 3.7
Incomplete spot measurements of phytoplankton and zooplankton biomass in Qinghai Lake, 1990 and 1992. Stations I–VI are as in Table 3.6, and stations 1–5 are in South Bay (section 3.1.2). ‘Small’ zooplankters (protozoans and rotifers) were from 55-μm net samples and ‘large’ zooplankters (crustaceans) were from 350-μm net samples (section 3.1.5). Numbers in normal font are individuals per litre; italic numbers are mg dry weight per litre except for those in 1992, which are 10^-4 mg dry weight per litre. The depths 5, 10 and 15 m are the depths to which the net was lowered before hauling to the surface.

Abundant growths of the green alga Cladophora cover most of the lake bottom up to 22 m depth in summer. Cladophora occurs in the summer plankton, and persists as a benthic species in other seasons. In 1961–63 the benthic algal biomass generally exceeded 50 g m^-2 wet weight and attained values as high as 2360 g m^-2, particularly in near-shore areas. Dense mats of algae may form long, deep wind-rows along the shore; these were extremely abundant near Bird Island in September 1992. The trawler captains have reported that mats of Cladophora cause problems for fishing in large areas of the lake, and are a primary reason why trawling must be concentrated in the South Bay and near the Buna mouth. Handlers at the Fish Factory often must discard tangled algal mats brought in with the catch.

The Academia Sinica survey showed that the principal zooplankton species include representatives of 17 genera, particularly the copepod Arctodiaptomus, the cladoceran Moina, the rotifers Brachionus and Hexarthra (Pedalia) and the protozoans Vorticella and Strombidium (see Appendix III). Protozoans contribute >80% of the biomass, expressed as numbers of individuals, but presumably are less significant as food for juvenile fish than crustaceans and rotifers. The nauplii (larval stages) of copepods are especially abundant in spring. Pronounced diel variations occurred in the vertical distributions of both phytoplankton and zooplankton, but there was no clear pattern of reactions to environmental factors like light intensity, temperature or weather conditions.

3.11 Lake Fauna

Appendix III shows a list of fauna recorded from Qinghai Lake, compiled from information in the bibliography (Appendix I), including the Academia Sinica (1979) survey, and from unpublished information contributed by staff of the Northwest Plateau Institute of Biology, Xining.