|Other Abstract||The Yangtze floodplain is one of the most important wetlands in the world. Many lakes in the region have suffered from fishery over-exploitation and man-made eutrophication for decades. To solve these problems, it is necessary to establish a regional-scale quantitative platform of lake ecosystem management. Traditional limnological researches were carried out in individual waters through small- to medium-scale analyses. Although these studies contributed greatly to our understanding of ecological mechanisms, the results, especially quantitative relationships are difficult to be extrapolated to real ecosystems or other waters, and the strategy is also insufficient to reveal macroecological patterns of complex systems. As a new approach, predictive limnology aims at finding general rules through large-scale comparative studies. In the present dissertation, predictive limnological researches on mid-lower Yangtze shallow lakes were carried out systematically in 46 small- to medium-sized lakes. The main results are as follows:
1. Suitable transformations of various limnological parameters have been recommended. Generally, data used in statistical analyses must follow a normal distribution and those with non-normal distributions must be normalized through transformations. The results showed that the frequency distributions of pH and water temperature are normal, and those of mean depth and ratio of Secchi depth to water depth are approximately normal; no transformation is needed for these four parameters. The frequency distributions of relative density and biomass of gastropods are U-shaped; so that they are difficult to be normalized through transformation. The coefficient of development of lake volume is right-skewed and can be normalized through x2 and ex transformations. All the other parameters are left-skewed and can be normalized through log10(x), x0.5 and x0.1 transformations. Since the data set is large and typical, the results are expected to be applied to quantitative limnological works on Yangtze waters.
2. Key-time models of submersed macrophyte biomass have been established. To recover the submersed vegetation in Yangtze lakes, the creation of models to predict growing tendency of vegetation is necessary. Annual investigations showed that the ratio of Secchi depth to water depth is the most important factor regulating the biomass of submersed macrophytes. Further analyses indicated that the months from March to June are not only the actively growing season for most macrophytes, but also the key time the key factor acts. Accordingly, a series of key-time models of submersed macrophyte biomass are generated using the ratio of Secchi depth to water depth during the key time as the driving variables (p<0.001):
Mar.: BMac = - 3149 + 4854.6 ZSD/ZM R 2= 0.75 n = 15
Apr.: BMac = - 3396 + 7298.6 ZSD/ZM R2 = 0.76 n = 16
May: BMac = - 3490 + 6380.6 ZSD/ZM R2 = 0.77 n = 17
Jun.: BMac = - 3536 + 7900.6 ZSD/ZM R2 = 0.69 n = 18
Mar.-Jun.: BMac = - 3931 + 7072.9 ZSD/ZM R2 = 0.81 n = 18
Where, BMac (wet weight, g/m2) is annual biomass of submersed macrophytes, ZSD/ZM is the ratio of Secchi depth to water depth during the key period. According to the models, the ratio of Secchi depth to water depth should reach over 0.66, 0.47, 0.55 and 0.45 respectively in the four months during March-June to enable a normal growth of submersed macrophytes. These models have high predictive abilities and thus provide a quantitative tool for recovery of submersed macrophytes through regulation of water level.
3. Predictive models of epiphytic gastropods on submersed macrophytes have been established. Epiphytic gastropods are an important group in shallow lakes. However, little has been done previously in China, especially in the aspect of modeling. Annual investigations showed that the community is characterized by the constitution of small individuals, mainly represented by Radix spp., Planorbidae spp., Alocinma longicornis and Parafossarulus striatulus. The mean body weight was 0.05 g/ind and the dominant (99%) size was less than 0.2 g/ind. The average density and biomass were 417±160 ind/m2 and 18.05±7.43 g/m2, with maxima around August. Biomass of submersed macrophytes was found to be the important factor affecting epiphytic gastropods. Accordingly, a series of annual and seasonal predictive models yielding high predictive abilities are generated (Annual: species number, R2=0.25-0.36, n=20, p=0.02-0.005; density and biomass, R2=0.48-0.69, n=20, p<0.001. Seasonal: density and biomass, R2=0.49-0.94, n=12-20, p<0.002). Further analyses found that pulmonates prefer the macrophytes which can spread their terminal parts on the water surface, while prosobranchs prefer the macrophytes which are entirely submersed. The models are expected to benefit rational utilization and protection of gastropod resources.
4. Preliminary predictive models of benthos standing crops have been established. The results showed that oligochaetes of Yangtze lakes were 403 ± 225 ind/m2 in density and 1.12 ± 0.39 g/m2 in biomass, gastropods were 82 ± 20 ind/m2 and 26.38 ± 3.99 g/m2, chironomids were 356 ± 62 ind/m2 and 1.86 ± 0.58 g/m2, the total were 847 ± 248 ind/m2 and 29.41 ± 3.97 g/m2. Water depth, Secchi depth, water temperature, phytoplankton chlorophyll a and submersed macrophyte biomass were found to be the important factors affecting the standing crops of benthic animals. Accordingly, a series of predictive models of benthos are generated using these factors as the driving variables.
5. Maximum yield models and an optimal-stocking model of Chinese mitten crab have been established. Overstocking of Chinese mitten crab in lakes has resulted in exhaustion of resources and deterioration of water quality; thus it is necessary to establish a method to estimate the optimal stocking rate of crab juveniles. Annual analyses indicated that submersed macrophyte biomass is the key factor affecting the crab yield. For convenient applications in crab culture, the ratio of Secchi depth to mean depth, a parameter easy to be accurately measured and close to both submersed macrophyte biomass and crab yield is selected as the driving variable. The ratio of Secchi depth to mean depth during the crab planting season (Dec.-May) is used as the driving variable to generate a series of maximal yield (CYMax, kg/ha) models (n=18, p<0.001):
Dec.-Jan. (1 month): CYMax = - 21.37 + 81.16 ZSD/ZM R2 = 0.61
Mar. (1 month): CYMax = - 1.22 + 63.9 ZSD/ZM R2 = 0.49
Apr. (1 month): CYMax = - 7.5 + 94.56 ZSD/ZM R2 = 0.64
May (1 month): CYMax = - 24.58 + 104.31 ZSD/ZM R2 = 0.70
Dec.-Mar. (2 months): CYMax = - 25.17 + 85.26 ZSD/ZM R2 = 0.65
Mar.-Apr. (2 months): CYMax = - 16.01 + 91.29 ZSD/ZM R2 = 0.71
Apr.-May (2 months): CYMax = - 23.21 + 109.26 ZSD/ZM R2 = 0.74
Dec.- Apr. (3 months): CYMax = - 32.37 + 103.18 ZSD/ZM R2 = 0.77
Mar.- May (3 months): CYMax = - 24.72 + 102.52 ZSD/ZM R2 = 0.75
Dec.-May (4 months): CYMax = - 36.60 + 110.69 ZSD/ZM R2 = 0.81
Validation based on an independent data set indicated that these models have high predictive powers, with a mean accuracy of 70%.
Based on the theory of MSY (Maximum Sustainable Yield), in combination with body-weight (BW, g/ind) and recapture rate (RR, %) of adult crabs, a general optimal-stocking model is formulated: SROpt=(1000CYMax×50%)/(BW•RR). According to the models, the optimal stocking rates of crab juveniles (yearlings, about 10±5 g/ind) in macrophytic lakes are generally 700±60 ind/ha. The models provide a simple and practical tool for rational crab culture.
6. The traditional viewpoint to use TN/TP ratio as an index to identify phytoplankton as nitrogen- or phosphorus-limited has been disproved. Traditionally, growth of phytoplankton in lakes have been regarded as limited by total phosphorus if TN/TP was relatively large, limited by total nitrogen if TN/TP was relatively small and co-limited by total nitrogen and total phosphorus when TN/TP was intermediate. However, this viewpoint has never been proved by strict statistical tests. Serial regression analyses in the present research indicated that total phosphorus was the primary affecting factor and total nitrogen the second affecting factor for both annual and summer phytoplankton chlorophyll a. In separate nutrient-chlorophyll a regression analyses for lakes of different TN/TP ratios, total phosphorus is also superior to total nitrogen in explaining the variation of chlorophyll a at all particular TN/TP ranges and over the entire TN/TP spectrum. Further analyses found out that chlorophyll a varied regardless of the changes of TN for a given amount of TP (annual, p=0.33; summer, p=0.81), but increased rapidly with an increase of TP for a given amount of TN (annual and summer, p<0.001). These results suggest that TN/TP ratio is inappropriate as an index to identify limiting nutrients and that TP is the primary nutrient limiting phytoplankton over the entire TN/TP spectrum. Natural phytoplankton communities are ones of multiple-species. Optimal N/P ratios vary greatly from species to species (ranging from 4.1-133.3 among various freshwater species). Obviously, it is almost impossible to set a specific “cut-off” ratio to identify a limiting nutrient(s) for a multiple-species community. In lakes, biological and lightning nitrogen fixation may correct for nitrogen deficiency. On the other hand, relevant mechanism is lack for phosphorus, resulting in the wide occurrence of phosphorus limitation. This study pinpoints the misleading of TN/TP ratio in lake management and highlights the vital importance of phosphorus abatement in eutrophication control.
7. Planktivorous fishes have been found to fail to control total phytoplankton. Previously, there has long been a controversy on whether planktivorous fishes could effectively control total phytoplankton. The controversy is mainly due to the fact that that all the experiments were too small in spatio-temporal scales and difficult to reflect the real ecosystems. In the present research, to test the effects of planktivorous fishes on total phytoplankton (measured as chlorophyll a), more than 70 sets of data were used and divided into two groups with fish yields (silver and bighead carp, the species native to China) greater than and less than 100 kg/ha. The results showed that, lakes with yields of planktivorous fish greater than 100 kg/ha have significantly higher phytoplankton chlorophyll a (p<0.001) and lower Secchi depth (p<0.001) than those with yields less than 100 kg/ha. Total phosphorus- chlorophyll a (slope difference, t=0.40; intercept difference, t=0.52; n=70，p>0.50, t0.50, 70=0.68) and total phosphorus-Secchi depth (slope difference, t=0.67; intercept difference, t=0.22; n=78, p>0.50, t0.50, 80=0.68) relationships are not significantly different between lakes with yields greater than or less than 100 kg/ha. These results indicated that the fish fail to decrease chlorophyll a yield, and also fail to enhance Secchi depth. The main reason is the selective filter-feeding on larger food particles. Small-sized species would develop quickly when their competitors (large-sized phytoplankton) and grazers (zooplankton) are eliminated. Therefore, silver carp and bighead carp are not recommended as a biotic agent for phytoplankton control in lake management if the goal is to control the total phytoplankton and to enhance water quality. For lake management concerning total phytoplankton control, nutrient abatement should be carried out first, and then try to recover submersed macrophytes and ecosystem integrity.
8. The regime multiplicity of shallow lakes has been proved; the primary factor and its thresholds triggering the regime shifts have been determined. Researches on multiplicity of lake regimes and mechanism triggering regime shift are important to lake management. They may help to prevent catastrophic events and define targets in lake management. However, in previous studies, the multiplicity of lake regimes was analyzed mainly by the means of experiments and models. Little evidence has been obtained from field data. In the present research, large-scale field data were used. It is proved that there are at least two ecosystem regimes in Yangtze lakes: one of clear water with submersed vegetation-dominance and one of turbid water with phytoplankton-dominance. Total phosphorus is determined as the primary factor triggering the regime shifts. The threshold of total phosphorus is determined as 70-100 mg/m3 for the clear-turbid shift, and 20-30 mg/m3 for the turbid-clear shift. The results provide quantitative references to define targets of nutrient control in Yangtze lakes. They are also expected to make a contribution to the theory of multiple regimes.
9. A special approach to practical models has been developed and the new concept “key-time model” has been defined. It is suggested that modelers should first determine regulating factors of target variables, and then seek for driving variables which are correlated closely with the regulating factors and easy to be accurately measured. Models generated like this are expected to have higher predictive powers and to be more practical. Key-time models are based on driving variables during the functioning period. They have higher predictive capacity than traditional synchronic models in general, and enable us to predict annual tendency in advance; thus are of greater value in application.
In summary, in the present dissertation, a series of simple and practical holistic empirical models have been established; several important hypotheses have been tested; some explicit and quantitative suggestions for lake management have been given; a new approach of predictive modeling has been developed. To create an expert management system of inland-waters in China, predictive limnological work will be carried out more extensively and intensively in the Yangtze Basin and other regions in the future.|