The Yangtze River Floodplain: Threats and Rehabilitation

Hong-Zhu Wang*, Xue-Qin Liu, and Hai-Jun Wang State Key Laboratory of Freshwater Ecology and Biotechnology Institute of Hydrobiology, Chinese Academy of Sciences No. 7 Donghu South Road, Wuchang District, Wuhan 430072, China

Abstract.—The Yangtze (Changjiang) River floodplain is one of the most important ecosystems in China, as well as in the world, but is seriously threatened by multiple factors. Thus, it is crucial and urgent to rehabilitate the river floodplain. This paper reviews ecological studies conducted on the Yangtze River floodplain and presents suggestions for conservation and rehabilitation. First, the Yangtze River system is briefly introduced. Formed 23 million years ago, the Yangtze River is ca. 6,300 km in length with a mean annual runoff of 9.6 × 1011 m3. Thousands of floodplain lakes are distributed along the mid-lower Yangtze River, and the total area remains 15,770 km2 at present. Such a river-lake complex ecosystem holds a unique and diverse biota, with ca. 400 hydrophytes and hygrophytes, ca. 170 mollusks, ca. 200 fishes, ca. 400 water birds, and endangered dolphins and porpoises. Second, main threats to the Yangtze River floodplain ecosystem are identified: (1) habitat loss, including river channelization, sharp shrinkage of lake area (ca. 10,000 km2 since the 1950s), degradation of lakeshore zones, and sand overmining; (2) alternations of hydrological regimes, including construction of ca. 47,000 reservoirs and disconnection of most lakes from the main stem; (3) water pollution, including eutrophication, heavy metals, and organic pollutants; and (4) overexploitation of biological resources, including overfishing and intensive pen culture. Third, effects of river–lake disconnection on lake ecosystems are summarized on the basis of our studies in the past 20 years. It was found that (1) disconnection is one of the main causes of lake eutrophication; (2) species diversity, biomass, and production of macrophytes and macrobenthos reach maxima at some levels of intermediate river connectivity; (3) disconnection greatly reduces fish species richness of each habitat guild, and natural fish larvae is severely depleted; and (4) disconnection simplifies macroinvertebrate food web structure, and the trophic basis of the simplified food web is more heavily dependent on detritus in disconnected. Last, conservation strategies are proposed. Since the Yangtze River floodplain is a huge integrated system, the biodiversity conservation must be conducted on the whole basin scale. By establishing species–area models of fishes, the minimum protected area of Yangtze-connected lakes is estimated to be ca. 14,400 km2. It means that at least 8,900 km2 of disconnected lakes should be reconnected with the Yangtze main stem, and ecohydrological operation of dams and sluices is the feasible approach. Based upon our preliminary studies on environmental flow requirements, the following measures are suggested: (1) lower water levels during spring to improve germination of macrophytes, and control rising rates of water levels during spring–summer to ensure development of macrophytes; and (2)

* Corresponding author: [email protected]

1 2 wang et al.

open sluice gates to restore migration routes for juveniles migrating into lakes during April–September and for adults migrating back to the Yangtze main stem during November–December.

Introduction area of more than 1,000 km2, and 22 rivers, each with a catchment area of more than The Yangtze River floodplain is one of the 10,000 km2. Numerous lakes are distributed most important ecosystems in China, as on the plateaus and floodplains. Plateau well as in the world, but is facing multiple lakes in the upper basin are mainly freshwa- stressors, such as habitat loss, alternations of ter lakes, with a few saltwater ones. Flood- hydrological regimes, pollution, and overex- plain lakes in the middle and lower basins ploitation. Therefore, it is crucial and urgent are all freshwater ones (Yu and Lu 2005). to rehabilitate the river floodplain. In this The mean annual runoff of the Yangtze paper, we present a review of ecology and River is 9.6 × 1011 m3, being one-third of the conservation of the Yangtze River floodplain. total amount discharged into the sea by Chi- First, we briefly introduce the Yangtze River nese rivers and ranking forth in the world. system. Second, we identify main ecological Approximately 46% of the Yangtze runoff threats. Third, we summarize the effects of is from the upper basin, 18% from the Lake river–lake disconnection on lake ecosystems. Dongtinghu drainage system, and 15% from Last, we propose holistic strategies for con- the Lake Poyanghu system. Runoff of the servation and rehabilitation. main stem and tributaries is markedly different between flood and nonflood seasons The History and Current Status of in a year. The runoff of the main stem in the Yangtze River Floodplain the flood season (May–October) is 79–82% of the annual total in the upper basin, and Introduction to the Yangtze River Basin and 71–79% in the mid to lower basins (Yu and Its Geological History Lu 2005). The Yangtze River is the longest river in In the Early Cretaceous period, the Yan- Asia and the third longest in the world. shan Movement turned the entire Yangtze Originating from Geladandong Peak, the River basin into land. However, the Paleo- main peak of the Tanggula Mountains on Yangtze had not yet born until the Paleo- Qinghai–Tibet Plateau, the Yangtze River gene. From the Late Paleogene to the Early flows ca. 6,300 km from west to east and Neogene, the uplift of the Qinghai–Tibet enters the East China Sea. The total catch- Plateau altered the air circulation of lower ment area is 1.8 × 106 km2, being one-fifth atmosphere, turned the trade wind into the of China’s land territory. The terrain slopes seasonal wind, and changed the climate downward stair-like from west to east, from hot and dry to warm and humid. Since with a total fall of ca. 5,400 m. The upper the precipitation and runoff sharply in- reaches are the course above Yichang, being creased, the ancient Yangtze River started 4,504 km in length and 1 × 106 km2 in catch- to develop in several regions. The water ment area; the middle reaches are between system of the upper reaches, including the Yichang and Hukou, being 955 km in length Jinshajiang, Yalongjiang and Chuanjiang and 6.8 × 105 km2 in catchment area; and the (Sichuan River) paleorivers, ran south into lower reaches start from Hukou, being 938 the South China Sea, and the rivers below km in length and 1.2 × 105 km2 in catchment Yichang drained into the Jianghan basin area (Yu and Lu 2005). and others (Yu and Lu 2005). With regards There are more than 10,000 tributaries to the age of the Yangtze River system, in the Yangtze River system. Among them, Zheng et al. (2013) argued that the connec- there are 437 rivers, each with a catchment tion through the Three Gorges must post- status and conservation of yangtze cetaceans 3 date 36.5 Ma (million years ago) and a river, Distributions of Rivers and Lakes in the much like the modern one, was in existence Mid to Lower Yangtze Plains before 23 Ma. Downstream from the Three Gorges, the During the Late Pleistocene (23–15 ka BP Yangtze River runs east through the Lake [thousand years before present]), the climate Dongtinghu Graben basin, Lake Poyanghu was dry and cold, and the level of the East Graben basin, lower Yangtze Geosyncline China Sea was 130–160 m lower than the pres- belt, and Taihu–Wusong Graben basins, ent level. The bed of the Yangtze River was forming a series of fluvial-lacustrine plains intensely downcut, and headward erosion along the mid to lower reaches by sedimen- gradually moved up to Yichang, forming a tation. The plains (108–122°E, 24–34°N) lie deep narrow valley in the mid to lower reach- east–west for 1,800 km with a total area of es; compared to the present level, the riverbed ca. 1.6 × 105 km2. The terrain is flat and low, was 5–25 m lower and the mean water level with altitudes mostly ca. 50 m above sea was 20–40 m lower. In the Early Holocene (13– level (ASL). Plains along the middle reaches 12 ka BP), the climate became warmer, and the consist of the Jianghan plain, Lake Dongtin- level of the East China Sea was uplifted to an ghu plain, Edong plain, and Lake Poyanghu altitude 50–100 m below the present level. In plain; those along the lower reaches comprise the Mid-Holocene (7.5–4 ka BP), temperature the Chaohu–Wanjiang plain and the Yangtze reached the highest level of the postglacial River delta among the Jiangsu and Zhejiang period; the mean annual temperature in the provinces and Shanghai (Figure 1; Zhao and lower Yangtze basin was 3–4°C higher than Chen 1999). the present one, and the sea level was 2–3 m The mid to lower Yangtze main stem higher. At that time, the Yangtze estuary was is 1,893 km, with a sinuosity of ca. 1.6 (Fig- around Zhenjiang and Yangzhou. Due to ure 1). The main channel pattern is braided large areas of forests and lakes, the amplitude reaches, being 70% of the mid to lower river of water level fluctuation in the river was rela- in length. The river course is characterized by tively small (IGSNRRCAS et al. 1985; Yu and alternation of broad braided reaches and nar- Lu 2005). row single channels, resembling lotus roots. In the Late Holocene (4 ka BP to now), At bank-full level, the mean river width (B) is the climate fluctuated between cold-dry 1,168–3,332 m and the mean depth (H) is 13– and hot-wet, the amplitude of temperature 20.4 m, with B0.5/H of 1.88–4.45, and the max- change was ca. 3.5°C (Zhu 1972), and the imum depth of the thalweg reaches 90.5 m. sea level was ±3 m around the present. Since The longitudinal profile of the mid to lower the Western Jin Dynasty (265–317), large course has a total fall of 56.4 m, with a mean populations in the north were compelled slope of 0.34 (0.05–0.85) × 10–4 (IGSNRRCAS to migrate to the south due to turmoil of et al. 1985). wars, and they deforested for farming and Numerous lakes are distributed along reclaimed land from lakes. Due to climate the mid to lower Yangtze River (Figure 1), change and aggravated soil water erosion, with a total area of 15,770 km2. They are di- sand bars emerged in the river course and vided into the following lake clusters: (1) the Yangtze estuary kept expanding fur- The Jianghan–Dongtinghu plains lakes, dis- ther east from Yangzhou, reaching some- tributed between Zhijiang and Wuxue. In where near the present estuary in the Tang the Mid-Holocene, this region was the vast and Song dynasties (618–1279). For reasons waters of the paleo-Yunmeng Lake. After- such as the shifting of the main stem to the wards, great quantities of sand brought by right and the construction of left banks, the Yangtze River and its tributaries were lakes along the left bank shrank gradu- deposited in the lake. As a consequence, land ally while lakes along the right expanded areas expanded while water areas shrank. (IGSNRRCAS et al. 1985). The lake was thus divided into thousands of 4 wang et al.

Figure 1. The mid-lower Yangtze River basin. smaller lakes. At present, there remain more Chaohu. (2) Lakes formed in abandoned than 600 lakes, including Lake Dongtinghu, channels. One type is formed by the main- Lake Honghu, and Lake Liangzihu. (2) The stem migration, including the oxbow lakes Gan-Wan lakes, distributed between Wuxue along the middle reaches and the Yangtze- and Datong. This group includes many lakes side lakes between Huangshi and Datong. in eastern Hubei, southern Anhui, and the Another type is irregular long lakes formed northern Jiangxi provinces, such as Lake by obstruction of tributaries. (3) Lateral levee Poyanghu, Lake Longganhu, Lake Bohu, and lakes, formed by overflowing narrow -low Lake Caizihu. (3) The Su-Wan lakes, distrib- lands or valleys along floodplain boundaries. uted between Datong and Maoshan, includ- The outlets were blocked by sedimentation. ing Lake Chaohu, Lake Nianyihu, Lake Shi- Typical ones are Lake Luhu and Lake Liang- jiuhu, and Lake Guchenghu. (4) The Taihu zihu around Wuhan. (4) Scour lakes, deep plain lakes, distributed in the Yangtze River pools scoured by floods after banks burst. delta, including more than 200 lakes (Zhao (5) Lakes formed by river action on lagoons, and Chen 1999). such as Lake Taihu. The origin of Yangtze lakes is roughly classified into five types (Rao 1956; Wang Climate and Hydrologic Features of the and Dou 1998): (1) Lakes formed by sedimen- Mid to Lower Yangtze River Floodplain tation in graben/depression basins. Since The mid to lower Yangtze basin mostly be- sedimentation rates are higher than subsid- longs to the subtropical monsoon zone. The ence rates, the lakes become shallower grad- annual mean temperature is 13–20°C and the ually. Typical ones are the Jianghan lakes, annual accumulated temperature (≥10°C) is Lake Dongtinghu, Lake Poyanghu, and Lake status and conservation of yangtze cetaceans 5

4,000–6,500°C. The mean temperature of the ed lakes only during the high-water level pe- coldest month (January) is above 0°C, be- riod, with moderate fluctuation amplitude. ing 0–2°C in areas north of the main stem, Water levels are low and stable during the 2–10°C in areas south of the main stem, and low-water level period. The third WLF type 10–12°C in the Nanling mountainous areas. is reservoir-like WLF. It mainly includes The freezing duration is short, and the frost- Yangtze-disconnected lakes, with small fluc- free season lasts for ca. 285 d. The annual tuation amplitude. Before the flood season, precipitation is 800–1,600 mm. It decreases lakes are drained to the lowest water lev- from southeast to northwest, being 1,200– els for flood control. After the flood season, 1,800 mm in east hill areas, ca. 1,500 mm sluice gates are shut to store water and lakes in mountainous areas south of the Yangtze are kept at high water levels. main stem, and 1,000–1,200 mm in the Yang- tze plains. Seasonally, precipitation ranks Environments and Biological Resources of first in summer, being 40% of the annual to- the Yangtze River Floodplain tal in June–August; second in spring; third The Yangtze floodplain lakes are character- in fall; and least in winter (but also >10%) ized by their shallow, flat, and large basins. (Zhao and Chen 1999). The mean water depth of these lakes is gen- In the mid to lower Yangtze main stem, erally ca. 2 m and rarely over 4 m. There are the multiyear mean discharge is from 13,900 18 lakes each with an area larger than 100 3 3 m /s at Yichang to 28,700 m /s at Datong, km2. Because of these characteristics, interac- and the mean annual runoff is from 4.38 × tion between water and sediments happens 11 3 11 3 10 m at Yichang to 9.60 × 10 m in the estu- frequently in the lakes under the influence ary. The runoff in May–October is more than of monsoon climate. Moreover, the lakes 70% of the annual total, with a peak in July. are characterized by high levels of nutrients. The multiyear mean water level is from 43.8 Thus, the transparency is quite low (Table 2). m ASL at Yichang to 8.7 m ASL at Datong. In the main stem, total nitrogen was lower, The water level is highest in July–September but total phosphorus was higher (Table 2). and lowest in December–February, with an- The nutrient levels of the lakes were nual fluctuation amplitude of 13.5–18.4 m. found to be high (ca. 50 μg/L) back in his- –4 The water table slope is (0.193–0.579) × 10 tory, according to a palaeolimnological re- in the middle main stem and (0.097–0.0203) search (Yang et al. 2006). With a combination –4 × 10 in the lower. The sediment content of high nutrients and good water-heat con- of water is moderate. The multiyear mean ditions, a unique and diverse biota has been 3 sediment concentration is from 1.14 kg/m evolved in the Yangtze River and lakes com- 3 at Yichang to 0.48 kg/m at Datong, and the plex ecosystems. There are ca. 400 recorded mean annual sediment runoff is from 5.0 × species of hydrophytes and hygrophytes 8 8 10 metric tons (mt) at Yichang to 4.3 × 10 mt (Zhang 2013), 167 mollusk species with 92 at Datong (Yu and Lu 2005). endemic to China (Shu et al. 2009), ca. 200 As a consequence of human disturbance, fish species with 50% endemic to China (Yu hydrological regimes of the Yangtze flood- et al. 2005; Liu and Wang 2010), ca. 50 reptile plain lakes have been altered to various ex- species, ca. 400 water bird species, and more tents. According to water level fluctuations than 60 mammal species, including endan- (WLFs), three lake types can be identified gered river dolphin Neophocaena phocaenoides (Table 1). The first WLF type is quasi-natural. and finless dolphin Lipotes vexillifer (Fang et It is similar to the WLF of the Yangtze main al. 2006a). stem, with large fluctuation amplitude. This The abundances of plankton, macro- type includes all Yangtze-connected lakes phytes, and macroinvertebrates are rather and some disconnected ones. The second high in the lakes, but very low in the main WLF type is intermittent. This type includes stem (Table 2). Fishes are abundant in the lakes connected to rivers or Yangtze-connect- 6 wang et al.

1.92 1.17 6.40 2.17 1.18 18.30 water Annual ) turnover 3 6.58 1.67 0.99 (m 12.65 19.00 44.28 Lake 178.00 249.00 volume 2.13 2.82 3.50 3.40 6.98 2.46 1.26 (m) WLF 13.35 13.31 amplitude 2.20 6.20 6.00 3.96 3.50 3.77 2.60 (m) 23.50 29.19 depth Maximum

(m) 6.39 1.91 4.16 5.10 4.55 0.45 1.26 2.69 1.90 Mean depth

) 2 3,265 1,554 9,258 area 10,352 36,895 (km 257,000 162,000 Catchment )

36.7 14.0 78.5 344.4 304.3 769.6 (km 2,432.5 2,933.0 2,338.1 Lake area type Water level Quasi-natural Quasi-natural Reservoir-like Quasi-natural Intermittent Intermittent Quasi-natural Reservoir-like Reservoir-like

fluctuation (WLF)

Connectivity with Yangtze Connected Disconnected Disconnected Connected Seasonally connected Seasonally connected Disconnected Disconnected Disconnected Hydrological features of typical lakes on the Yangtze floodplain. (sublake of Poyanghu) (sublake of Poyanghu) Table 1. Lake Dongtinghu Honghu Liangzihu Poyanghu Dahuchi Shahuchi Shengjinhu Chaohu Taihu status and conservation of yangtze cetaceans 7

Table 2. Basic limnological characteristics of lakes and the Yangtze main stem in the mid to lower Yangtze River basin (reanalyzed from Wang et al. 1999, 2005, 2006, 2007, 2008, 2014; Feng 2005; Pan et al. 2009, 2011a, 2011b, 2014; Wang and Wang 2009; Zhao 2010). Ind = individuals; “–” = no data. Yangtze Lakes main stem Parameters Mean Median Min Max (mean) Secchi depth, m 1.1 0.9 0.3 3.3 0.41 Total nitrogen of water, mg/L 2.02 0.94 0.07 13.7 0.66 Total phosphorus of water, mg/L 0.11 0.04 0.005 0.76 0.13 Phytoplankton chlorophyll a, µg/L 16.5 4.4 0.8 139 1.69 Density of phytoplankton, 104 ind/L 947 361 16 6,615 – Biomass of phytoplankton, mg/L 6.3 2.3 0.2 35.3 – Density of zooplankton, ind/L 2,032 1,518 233 16,145 17.5 Biomass of zooplankton, mg/L 2.2 1.5 0.4 8.3 0.07 Biomass of submersed macrophytes, wet mass, g/m2 931 295 0 9,484 Rare Biomass of emergent macrophytes, wet mass, g/m2 406 82 0.4 1,728 – Density of macroinvertebrates, ind/m2 864 259 8 23,194 91 Biomass of macroinvertebrates (mollusks with shells), wet mass, g/m2 30.0 16.1 0.005 313 0.17 river floodplain, and the fish yield of natural tress 4.9 × 106 m2, the number of spur dikes lakes was 75–135 kg/ha (Wu and Jao 1958). 685, and the total length of longitudinal dikes 20 km. The total length of reinforced bank is Main Threats to the Yangtze River ca. 1,400 km. The second one is cut-off of me- Floodplain Ecosystem anders. In 1966–1972, three meanders were cut off from the lower Jingjiang in the middle Habitat Loss Yangtze. The river length was shortened by River channelization.—Embankments along 78 km, with sinuosity reduced from 2.83 to the mid to lower Yangtze River was initiated 1.93, and the slope was increased. As a re- between 345 and 356 during the Eastern Jin Dy- sult, the riverbed was severely eroded and nasty and began to take shape during the Ming the water level was lowered, reducing river and Qing dynasties (1368–1911). Discrete dikes water flowing into Lake Dongtinghu. The had been gradually connected together by that third one is blockage of secondary channels. time, and the total length was ca. 3,600 km, Since the 1970s, secondary channels in 7 of 41 forming a huge system (Wang 1959). Prior to braided reaches were blocked. After block- 1950, the river course was changed frequently age, flow regimes were simplified and main due to the weakness of embankments. Since channels were deepened. With regard to then, embankment engineering along the ecological effects of channelization along the Yangtze main stem has been conducted on a mid to lower Yangtze, no systematic studies large scale and the river regime has been sta- have been conducted. The probable impacts bilized, resulting in serious channelization. are habitat loss, environmental heterogeneity The work includes three major engineer- decrease, and biodiversity decline. ing projects (Yu and Lu 2005). The first one Sharp shrinkage of lake area.—Since the is dykes reinforcement, especially along the Western Jin Dynasty (265–317), large fluxes reaches with bank caving. Over more than 50 of northern Chinese have migrated to the years, the total volume of enrockments is 8.9 Yangtze basin many times. Due to land recla- × 107 m3, the total area of sunken fascine mat- mation from lakes and sedimentation caused by deforestation, the area of Yangtze lakes 8 wang et al. began to decrease (IGSNRRCAS et al. 1985). land reclamation, and the second is sedimen- Land reclamation reached a considerable tation. Historically, forest coverage of the scale in the Southern Song Dynasty (1127– Yangtze basin was 50–60%, but it decreased 1279), rapidly increased during the Ming to 30–40% in the 1950s and to 10% in 1986 Dynasty (1368–1644), and peaked in the mid- (cited from Wei et al. 2007). This resulted in dle of the Qing Dynasty (1644–1911) (Du et increased soil erosion, and the rate of sedi- al. 2011). In 1949, the total area of lakes was ment accumulation reached 1.6–3.4 mm/ 25, 828 km2 (CWRC 1999). From the 1950s to year in three lakes of the Jianghan plain, and 1970s, the reclaimed area sharply increased, 10–30 mm/year in Lake Dongtinghu (cited and the area of remaining lakes is only 15,770 from Du et al. 2011). The siltation in lakes km2 at present. further stimulated the reclamation process. In the Jianghan and Dongtinghu plains Degradation of lakeshore zones.—In many of the middle basin, the total lake area was Yangtze lakes, shore zones have degraded reduced by 5.2% between the 1930s and or even disappeared due to human activi- 1950s, by 46.4% between the 1950s and 1970s, ties such as land reclamation, urbanization, and by 17.5% between the 1970s and 2000 and road construction. Natural shorelines (Du et al. 2011; Figure 2). The major cause is have been artificially replaced by hard banks

Figure 2. Lakes from the Jianghan Plain and the Dongtinghu Plain in the 1930s, 1950s, 1970s, and 2000. (From Du et al. 2011.) status and conservation of yangtze cetaceans 9

(rocks and concrete). For example, in Lake from 1.5 to 0.1–0.5 m. Turbid water affects Chaohu along the lower reaches, stone bank plankton directly and inhibits plant growth. is 129.0 km long, cement revetment is 40.7 km Third, pollution of wastewater, oil, garbage, long, and both are 90% of the whole shoreline noise, light, and so forth from mining ships is length (Wang et al. 2012). Meanwhile, a vast produced, and probably release of sediment area of the riparian zone has become imper- pollution due to disturbance is enhanced. meable ground. Degradation of lakeshores Fourth, feeding, breeding, migration, com- results in deterioration of aquatic and terres- munication, and other activities of fishes, fin- trial vegetation along shorelines, diminution less porpoise, water birds, and so forth are of capabilities to intercept external pollution affected. Fifth, hydrogeomorphology and and to reduce internal loading, and loss of related processes are changed, resulting in lake-ecosystem integrity. bank collapse and decrease of lake and wet- Sand overmining.—Sand mining in the land area. Yangtze main stem began in the 1950s, and limited amounts of cobbles were collected Alternations of Hydrological Regimes from the Jinjiang reaches during low-water Cascade reservoirs in the main stem and periods. Since the 1980s, the amount and ex- tributaries.—By 2009, ca. 47,000 reservoirs tent of sand mining by machines increased had been built in the Yangtze basin. The to- gradually and the ship number reached thou- tal storage capacity was 2.5 × 1011 m3, and sands, but most of these activities were un- the usable capacity was 1.2 × 1011 m3, being authorized. Due to serious damage of sand 25% and 2% of the multiyear mean surface overexploitation, the State Council enacted runoff of the Yangtze River. Among them, a regulation on sand mining in the Yangtze there were 166 large reservoirs, with 38 in River, which has been in force since 2002. Up the upper basin, 114 in the middle basin, and to now, sand mining in the main stem has 14 in the lower basin (Chen 2012). Moreover, been controlled (Wu and Li 2010). However, many new reservoirs are being constructed large-scale sand mining moved quickly to or planned, and the South–North Water Di- Lake Poyanghu and Lake Dongtinghu, and version Projects are being implemented. their tributaries afterwards. Taking Lake Due to multiple factors, such as hy- Poyanghu as an example, mining ships in- draulic projects, water use for industry and creased from 140 in 2001 to 450 in 2005; after agriculture, climate changes, and land-use 2010, mining area enlarged from the north changes, hydrological regimes of the mid to the middle, and more than 90 additional to lower Yangtze main stem and floodplain ships were added (Cui et al. 2013). Recently, lakes have been greatly altered. Next, we take local governments began to issue relevant the changes after impoundment of the Three rules to control the serious situations of un- Gorges Reservoir (TGR) as an example. First, authorized sand mining. the mean annual runoff below the dam after Sand overmining should have very seri- the impoundment was reduced by 6–10% ous impacts on river floodplain ecosystems, during 2003–2007 (CAE 2010). The water re- but no systematic studies have been conduct- lease from TGR was decreased greatly from ed. According to some reports (Zhang and mid-September to November to store water, Huang 2008; Wu and Cui 2008; Ma and Xu but it increased from January to June; dur- 2013), the negative effects may include five ing the major flood season, it was reduced aspects. First, large amounts of sediment are only at risk of serious flooding (Wang and pumped out and habitats of benthic animals Chen 2010; Zhang et al. 2013). Second, the and aquatic plants are destroyed. Second, mean annual sediment discharge below the suspended substance dramatically increases. dam after the impoundment was greatly re- For example, in northern Lake Poyanghu, duced by 63–86% during 2003–2007, and the 3 suspended substance reached 0.35 kg/m , sediment released from TGR became much and Secchi depth in the summer dropped 10 wang et al. finer. This resulted in riverbed erosion, even invertebrates such as bivalves by drought down to the Datong reach in the lower ba- (Xu and Chen 2013; Zhang 2013; Zhang et al. sin, and the water level at equal discharge 2014). Since vegetation can provide important during the flood season was decreased by habitats and food bases, wildlife resources, 0.1–1 m. The most serious erosion occurred including water birds and fishes, were affect- in the Yichang–Chenglingji reach, with scour ed seriously (CAE 2010). Third, the lowering depth in the thalweg up to more than 10 m. of flood peaks and water temperature from The amount of sediment flowing into Lake spring to early summer retards fish spawn- Dongtinghu from the Yangtze main stem was ing, and fish resources declines seriously. Af- greatly reduced, and the mean annual sedi- ter the impoundment of TGR, the initiation mentation rate in 2003–2006 was reduced by of the spawning season was delayed ca. 1 90.7%, thus prolonging the lifetime of Lake month, early growth and development was Dongtinghu (CAE 2010). Third, water stor- suppressed, and the eggs and larvae were age in TGR and riverbed scour downstream reduced by 90% (Xie et al. 2007; Zhang et al. changed water levels of the main stem and 2011). Also, the earlier drop of water levels in Yangtze-connected lakes and their fluctua- autumn may disturb fish migration (Ru and tion rhythms. It was estimated that the wa- Liu 2013). ter level in the middle reaches decreased River–lake disconnection.—From the mid- by 2–3 m while storing water and increased 1950s to 1960s, most lakes were disconnected by 0.6–0.8 m in January–March (Wang and from the Yangtze main stem by embank- Chen 2010). During 2003–2007, the water ments and sluice gates, leaving few connect- level of Lake Dongtinghu (at Lujiao station) ed lakes, including Lake Dongtinghu, Lake increased by 0.5–1 m in January–March but Poyanghu, and Lake Shijiuhu, at present. Af- deceased by 1.2–1.4 m in July–August and by ter disconnection, water level fluctuations in 1.5–2.2 m in October–November; the water the lakes became reservoir-like. Taking Lake level of Lake Poyanghu (at Hukou station) Chaohu (777 km2) as an example (Figure 3), increased by 0.9–1.4 m in January–March before damming the water level reached a but decreased by 0.1–0.2 m in July–August maximum in summer and dropped to a mini- and by 1.3–1.5 m in October–November (Shi mum in early spring, with fluctuation ampli- et al. 2011). Fourth, water temperature was tude of over 2 m; afterwards, the amplitude changed. During 2003–2008, the water tem- was only 1 m and water level dropped to a perature at Yichang gauging station of the minimum before the flood season. Ecologi- main stem increased by 0.5–2.3°C in Octo- cal effects of river–lake disconnection will be ber–January but decreased by 0.3–1.8°C in discussed in detail later in this paper. March–May (Li et al. 2013). Water pollution.—The discharge of waste- The alterations of hydrological regimes water and sewage in the Yangtze River basin exert great impacts on the river floodplain has been continuously increasing since the ecosystem along the mid to lower Yangtze 1970s. The total amount increased from 9.5 × River. First, the drop of water levels in the 109 mt/year in the late 1970s through 1.5 × Yangtze main stem during summer and au- 1010 mt/year in the late 1980s and 2.0 × 1010 tumn blocks the exchange of matter and en- mt/year in the mid to late 1990s to 4.4 × 1010 ergy with the floodplain, decreasing flood- mt/year in 2012 (Figure 4; MWR 2004–2013). plain area, habitat heterogeneity, and species More than 80% of the wastewater is dis- diversity (Poff and Hart 2002; Shi et al. 2011). charged into the mid to lower basins. Second, alternations of water level fluctua- Lake eutrophication is one of the severe tion rhythms affect wetland vegetation. The impacts of excessive wastewater and sew- rise of water levels in spring can retard ger- age. More than 40% of the lakes were in eu- mination of hygrophytes and emergent mac- trophic–hypertrophic states (Figure 5). Two rophytes, and the drop of water levels in au- of the four largest lakes herein, Lake Taihu tumn can kill submersed macrophytes, and and Lake Chaohu, have suffered very severe status and conservation of yangtze cetaceans 11

Figure 3. Alteration of water levels before (1956–1960) and after (2007–2011) building sluices in Lake Chaohu. (From Zhang et al. 2014.)

Figure 4. Amount of wastewater and sewage discharged in the Yangtze River basin (YRB) between 2003 and 2012 (108 metric tons) and its percentage of the total amount in China (YRB%). (Data from MWR 2004–2013.) 12 wang et al.

Figure 5. Trophic states of Yangtze lakes according to the fixed boundary classification system (OECD 2006). (From Wang 2007.) status and conservation of yangtze cetaceans 13 cyanobacteria blooms for decades (Zhu 1996; ture, submersed macrophytes declined from Lin 2003; Xie 2009). coverage of ca. 90% in 1991 to less than 5% Pollution of heavy metals and organic in 2002 and disappeared in the end; Secchi pollutants are among other impacts on the depth was only 1.1 m, being around half of lakes. In the main stem (Wuhan, Hubei), the original state. Under the stocking den- ca. 64% of sediment sites were polluted by sity of 1 kg/ha of crab larvae or juveniles, heavy metals to a medium level of ecologi- the density and production of macroinverte- cal hazard; contents of eight metal species brates were reduced by more than 60% (Xu et were higher than the background soil levels al. 2003). Besides, fish culture using chemical (Shen et al. 2008). In lakes (Wuhan, Hubei), fertilizer is also common, especially in small heavy metals pollution in the sediment was to medium-sized lakes, thus worsening eu- more severe. The contents of four out of eight trophication in this region. metal species were more than twice as much as the background levels, and the average Ecological Effects of River–Lake content of zinc was as high as 297 μg/g in Disconnection on Yangtze Lake Donghu, Lake Longyanghu, and Lake Moshuihu (Shen et al. 2008). Concentrations Floodplain Lakes and the of persistent organic pollutants were higher Mechanisms than the standards in many lakes (Wuhan); We focused upon this issue in the past 20 19 species of organo-chlorine pesticides were years, and the results are summarized as detected in Lake Honghu (Fang et al. 2006b). follows. Overexploitation of biological resources.— Biological resources in the Yangtze River Plankton floodplain have been seriously overexploit- The mean species number of phytoplankton ed. In the main stem and connected lakes, in Yangtze-connected lakes was 25% higher eletrofishing boats and fine-meshed (0.5–1 than that of disconnected lakes while the cm) fish traps are the main direct threats to total density of the former was 42% lower fish resources (Cao 2011). Due to overfishing (Wang and Wang 2009). The dominant taxon as well as reservoirs construction, river-lake of connected lakes was cryptophytes while disconnection, and pollution, fish yield in the that of disconnected lakes was cyanobacte- whole river-lake system has decreased from ria. Densities of cyanobacteria, chlorophytes, 4.3 × 105 to 1.0 × 105 mt/year between the and bacillariophytes in connected lakes were 1950s and 2000s (cited from Liu and Wang obviously higher than those of disconnected 2010). The famous Reeves Shad Tenualosa lakes, while those of cryptophytes and eugle- reevesii has disappeared from the Yangtze nophytes were in reverse. The mean species River; Tapertail Anchovy Coilia ectenes and number of zooplankton (rotifers, cladocer- pufferfish have become rare, and natural ans, and copepods) in connected lakes was populations of the four major domestic Chi- ca. 30% lower than that of disconnected lakes, nese carps have sharply decreased (Cao 2008, and densities of the total and each group in 2011). Meanwhile, mollusks and other organ- the former were much lower. isms have been also overused as human and Phytoplankton chlorophyll a (Chl a) of fish food or for other purposes. disconnected lakes (30 μg/L) ranked first, In disconnected lakes, pen culture has lentic area of connected lakes second (5.0 become the dominant fishery model in the μg/L), and lotic area of connected lakes third past 20 years, overstocking Chinese mitten (2.1 μg/L), while total nitrogen (TN) and to- crab Eriocheir sinensis in most cases. Although tal phosphorus (TP) of lake water were in dif- useful for production and management, pen ferent order (TN: 2.6 mg/L, 1.2 mg/L, and culture can have severe negative effects on 1.5 mg/L; TP: 0.2 mg/L, 0.06 mg/L, and 0.13 ecosystems. In Lake Biandantang (Hubei), mg/L) (Wang and Wang 2009). This means after years of continuous intensive crab cul- 14 wang et al. that the highest abundance of phytoplank- high concentrations of suspended sub- ton in disconnected lakes cannot be ex- stance, resulting in light limitation on algal plained by nutrient alone. A further analysis growth. The above analyses demonstrate showed that TP is the limiting factor of Chl that river–lake disconnection makes lake a in disconnected lakes as well as in lentic water more still, and then phytoplankton areas of connected lakes, but the slope of the propagate in large numbers. Similarly, due TP-Chl a regression in the former was about to stable environment and plentiful food of twice as great as that in the latter (Figure 6). phytoplankton, zooplankton develops well This disparity may mainly be attributed to in disconnected lakes. the different residence times. In lotic areas of connected lakes, the limiting factor of Chl Macrophytes a is water current. Chl a had a parabolic re- There was no significant difference in the lationship with surface velocity, reaching mean species number of the total as well as a vertex at the velocity of 0.12 m/s. When floating and submerged species between current is slow, Chl a tends to increase with Yangtze-connected and disconnected lakes velocity, due to the fact that higher flow is (Wang and Wang 2009). The species number closely linked to greater input of external of emergent macrophytes in disconnected nutrient sources. When current is fast, Chl a lakes was 40% higher than that of connected tends to decline, suggesting that current in- lakes while that of hygrophytes was 8% low- hibition has prevailed over nutrients effects er in the former. on growth of algae. Fast current may cause

Figure 6. Comparison of phosphorus–chlorophyll a regression lines from Yangtze-connected and disconnected lakes. TP = total phosphorus of lake water; Chl a = phytoplankton chlorophyll a; U = water velocity. (From Pan et al. 2009.) status and conservation of yangtze cetaceans 15

Hygrophytes develop very well in the fluctuation amplitude is decreased in- dis Yangtze-connected lakes, with coverage up connected lakes, thus resulting in decreased to 15% and biomass up to 1,500 g/m2, but coverage and biomass of hygrophytes and develop poorly in disconnected lakes. Emer- increased submerged macrophyte biomass. gent macrophytes can grow well in connect- After disconnection, the reduced rate of ed and disconnected lakes, with biomass up water-level change is also harmful to hygro- to 2,000 g/m2 but coverage lower than 10%. phytes development. If WLF type is changed Freely floating macrophytes occur more fre- as reservoir-like, this can lead to degrada- quently in disconnected lakes, but coverage tion of all vegetation. For example, vegeta- and biomass are very low. Floating-leaved tion coverage in Lake Chaohu has decreased macrophytes may cover large areas in the from 30% before disconnection to less than two lake types, but biomass is relatively low. 1% at present (Zhang et al. 2014). Second, Submerged macrophytes develop better in other habitat conditions are changed. In dis- disconnected lakes than in connected ones, connected lakes, water velocities decrease so with coverage up to 80% and 40%, respec- sediment concentrations of water decrease, tively, and biomass up to 2,600 and 1,200 g/ and the lake bottom is covered with more m2, respectively (Wang and Wang 2009). silt, thus favoring freely floating and -sub In terms of dominant species, hygro- merged macrophytes. phytes are dominated by the genus Carex in connected lakes and Cynodon dactylon in Macrobenthos disconnected lakes. Emergent macrophytes Compared with those of the Yangtze main are dominated by Phragmites australis and stem and disconnected lakes, macrozooben- Triarrhena lutarioriparia in connected lakes thos of the Yangtze-connected lakes are char- and by P. australis and Zizania latifolia in acterized by higher diversity (Wang et al. disconnected lakes. For submerged macro- 1999, 2007; Pan et al. 2011a). This pattern of phytes, the dominant species in connected species richness conforms to the theory that α lakes are Potamogeton malaianus, Hydrilla ver- diversity of macrobenthos in floodplain wa- ticillata, Ceratophyllum dermersum, and Val- ters reaches a maximum at an intermediate lisneria spiralis. Among them, P. malaianus level of connectivity (Amoros and Bornette prefers waters with larger water level fluc- 2002). Mollusks in connected lakes were es- tuation amplitude while the others prefer pecially species-rich, and the total number of areas with lower connectivity. The dominant species was about four times as high as that submerged species in disconnected lakes are in disconnected lakes; the total number of P. maackianus, H. verticillata, C. dermersum, insect species in connected lakes was about and V. spiralis. Among them, P. maackianus 1.5 times as high as that in connected lakes; dominates the climax community. The domi- the total number of oligochaete species and nant submerged species usually shifts from other animals in connected lakes was about P. malaianus to P. maackianus after river–lake twice as high as that in connected lakes. This disconnection. result seems attributable to habitat heteroge- There are two mechanisms for discon- neity within such a large lotic-to-lentic area. nection to change macrophytes. First, water For example, bottom structures of connected level fluctuations (WLFs) are altered (e.g., lakes range from silt to sand and to gravel, Figure 3). Water level fluctuation amplitude but in disconnected lakes, sediment is pri- is closely correlated with diversity and bio- marily with silt. Also, river–lake disconnec- mass of all macrophytes. Species richness tion can indirectly affect macrobenthos by maximizes at an intermittent level of fluctua- changing vegetation types. tion amplitude (ca. 5 m; Figure 7). Biomass River–lake disconnection affects stand- of hygrophytes increases with fluctuation ing crops of macrobenthos to a consider- amplitude while that of submerged macro- able extent (Pan et al. 2011a). Although total phytes decreases (Figure 7; Zhang 2013). The 16 wang et al.

Figure 7. Relationships between water level fluctuation amplitude, and species richness and biomass of macrophytes in Yangtze floodplain lakes. (From Zhang 2013.) densities had no much difference, connected leaves. Therefore, their abundance should lakes had abundant bivalve filters such as the increase due to more external organic matter genus Corbicula, which was more than eight brought by increased river inflow but may times denser than that in disconnected lakes. be inhibited in case of overdisturbance. At an Oligochaetes and insects were more abun- intermediate level of river connectivity, the dant in disconnected lakes, and the density substrate is hard and sandy in general (Fig- of each was 4.6 times and 4.5 times as high as ure 8C), and suspended organic particles are that of connected lakes, respectively. relatively abundant (Figure 8C). Thus, the Along the river connectivity gradient, substrate favors filterers (mainly bivalves). the change patterns of functional groups At a high level of river connectivity, the sub- of macrobenthos in Yangtze lakes could be strate becomes unstable and water is full of classified into three types (Pan et al. 2011a). inorganic particles so that filterers would First, density, biomass, and production of not be supported. The increase of preda- shredders, collector-filterers, and predators tors (e.g., the family Dytiscidae) at the initial show unimodal changes. Shredders (e.g., stage may be the result of reduction in mac- the genus Stictochironomus) feed on coarse rophytes because plants can provide refuges particulate organic matter, such as decaying for zoobenthos and impede predation. The status and conservation of yangtze cetaceans 17 decrease of predators at later stages seems riverine fishes by 71.7%, river–lake migra- to be caused by high turbidity and habi- tory fishes by 40.6%, and lake resident fishes tat instability. High turbidity may prevent by 25.4% (Figure 9A–D). The reduction in predators from seeing prey (Figure 8C). fish diversity after river-lake disconnection Second, density, biomass, and production should be attributed mainly to the blockage decrease when river connectivity increases. of migration routes, the loss of fluvial envi- This pattern is represented by collector-gath- ronments in which some species spend parts erers (mainly the families Tubificidae and of their life cycle, and decreased habitat het- Chironomidae). Collector-gatherers mainly erogeneity. consume fine particulate organic matter, so The resource of eggs and larvae of four their decrease probably is caused by the domestic Chinese carps in the Yangtze main fact that the organic-rich silt diminishes stem was estimated to be 1 × 1011 in 1964 (Yi with increasing river connectivity (Figure et al. 1988) but was reduced to only 1.7 × 1010 8C). Third, density decreases with increas- in 1981 (STSGDFCR 1982). River–lake dis- ing river connectivity, whereas biomass and connection is one of the main causes for the production change unimodally. Scrapers observed decline in fish larvae. exhibit this trend. The decrease of scrapers density could be ascribed to the reduction of Food Webs macrophytes, which small-sized gastropods As stated above, river–lake disconnection prefer (Figure 8C). The increase of biomass changes lake communities completely, thus and production at the initial stage could be affecting the food web structure. For exam- ascribed to the replacement of small-sized ple, macroinvertebrate food web structure epiphytic gastropods (mainly the family differs greatly between Yangtze-connected Bithyniidae) by large-sized bottom dwell- and disconnected lakes, and the food web of ers (mainly the genus Bellamya), and the the latter is much simplified (Figure 10; Liu decrease at later stages may result from the et al. 2006; Liu and Wang 2008). The num- increase of sand fractions. Because of insta- bers of total species, macroinvertebrate spe- bility and detritus shortage, sandy bottoms cies, functional groups, predators, and total in the Yangtze-connected lakes should not links are much higher in connected lakes be suitable for scrapers, which feed mainly than in disconnected ones. By contrast, con- on detritus there. In terms of total benthos, nectance and connectivity are much lower in density decreases with increasing river con- connected lakes, indicating that the species nectivity (Figure 8A), showing the same interaction strength is weaker in these lakes pattern with the dominant groups in den- than that of disconnected ones. Connected sity (i.e., scrapers and collector-gatherers). lakes have more heterogeneous habitats and Biomass and production change unimodally wider range of food resources, and this can (Figure 8B), showing the same pattern with decrease the interaction strength among spe- the dominant groups in biomass (i.e., scrap- cies. Also, the greater seasonal fluctuations of ers and collector-filters). physical environments such as water levels can decrease the frequency that animal spe- Fishes cies meet each other, and thus the interaction By comparing species–area relationships strength. Although the trophic basis of mac- of fishes in Yangtze-connected and discon- roinvertebrate food web is detritus in both nected lakes (Figure 9), our study shows that connected and disconnected lakes, the latter river-lake disconnection reduced the total is more heavily based on detritus (70%) than species number in lakes on the Yangtze River the former (40%). In terms of energy flow, the floodplain by 38.1% on average. In terms of total energy flowing through the food web habitat guilds in disconnected lakes, river– is much higher (by orders of magnitude) in sea migratory fishes was reduced by 87.5%, connected lakes. 18 wang et al.

Figure 8. Tendency diagrams to illustrate responses of (A) density of different functional feeding groups, (B) biomass and production of different functional feeding groups, and (C) environmental parameters to river connectivity in the Yangtze floodplain. (From Pan et al. 2011b.) status and conservation of yangtze cetaceans 19

1.2 1.4

1.0 1.2

0. 1.0

0. 0.

0.4 0.

0.2 0.4 0.0 0.2 0.0 0.5 1.0 1.5 2.02.5 3.03.5 4.0 0.00.5 1.01.5 2.02.5 3.03.5 4.0

2.0 1.

1. 1.4

1. 1.2

1.0 1. 0. 1. 0. 1.5 0.4 1.4 0.2 1.3 0.0 1.2 0.0 0.5 1.0 1.5 2.02.5 3.03.5 4.0 0.0 0.51.0 1.5 2.02.5 3.03.5 4.0

2.1

2.0

1.5

1.4 0.0 0.5 1.0 1.5 2.02.5 3.03.5 4.0

Figure 9. Species–area relationships of fishes in Yangtze-connected and disconnected lakes in the Yang- tze floodplain. (From Liu and Wang 2010.)

Mechanisms for Effects of River–Lake eutrophication, increasing the conversion Disconnection on Lake Ecosystems efficiency of nutrient into phytoplankton biomass to a large extent; (2) species diver- We have systematically revealed the rela- sity, biomass, and production of macro- tionships between river connectivity and phytes and macrobenthos reach maxima at lake ecosystems (Figure 11). The main find- some levels of intermediate river connec- ings are as follows: (1) river–lake discon- tivity; (3) river–lake disconnection greatly nection is one of the main causes of lake reduces fish species richness of each - habi 20 wang et al.

Figure 10. Comparison of niche overlap web (summary web) of Yangtze-connected Lake Dongtinghu with two disconnected lakes. IS = insect shredders, GS = gastropod scrapers, ACG = annelid collector-gatherers, ICG = insect collector-gatherers, BCF = bivalve collector-filterers, ICF = insect collector- filterers, SIP = small insect predators (mainly chironomids), and LIP = large insect predators (mainly odonates). The line thickness in each ellipse shows the dietary overlap of species within the functional feeding group. (From Liu and Wang 2008.) tat guild, and natural resource of fish lar- dietary overlaps, and trophic basis is more vae is severely depleted; and (4) river–lake dependent on detritus. disconnection simplifies macroinvertebrate River–lake disconnection disturbs the food web structure in disconnected lakes natural flow regimes, decreasing the species and increases the connectance and degree of diversity of lakes sharply and changing the status and conservation of yangtze cetaceans 21

Figure 11. Relationships between hydrological connectivity and different ecological groups in the Yangtze floodplain. community structure substantially. It has nisms (Ward 1997; Amoros and Bornette been revealed that natural flow regimes are 2002; Ward et al. 2002). First, block the ex- the main mechanisms by which river flood- change of water, sediment, nutrients, and or- plain ecosystems have quite rich species and ganic matter between the Yangtze main stem very high production (Mahoney and Rood and lakes, disturb periodicity of hydrological 1998; Amoros and Bornette 2002; Lytle and fluctuations, and hinder/delay growth and Poff 2004; Junk 2005; Rood et al. 2005). Over development of organisms. Free connection evolutionary time, all species have devel- is of particular importance to organisms with oped life history strategies adapting to pre- habitat shifts in life histories. Second, lessen dictability of water regimes, and thus hydro- hydrogeomorphological dynamics, and de- logical alternations will exert great effects on crease spatiotemporal heterogeneity of habi- any organisms. In summary, obstruction of tats. Hydrological processes create various river–lake hydrological connectivity affects landscape elements and environmental gra- Yangtze lake ecosystems via three mecha- dients, which are changing over time. Third, 22 wang et al. stop or weaken natural disturbance by hy- servation in this paper. According to our drological dynamics, which interrupts suc- studies and related literatures, we propose cessions of various communities and keeps the following two principles concerning bio- them in different successional stages, thus diversity conservation in the Yangtze River forming mosaics of community patches. floodplain. According to situations in the Yangtze The principle of holistic conservation floodplain, Figure 12 shows how river–lake Since the Yangtze River is a huge integrat- connection works on lake ecosystems: ed system formed before 23 Ma, biodiver- In winter (Figure 12A), water levels in sity conservation must be conducted on the the Yangtze main stem and lakes are the low- whole water system scale. The Yangtze River est and some areas are dry. Hygrophytes on floodplain covers a large geographical- re lakeshore and some submerged macrophytes gion and holds diverse subsystems such as in shallow waters begin to germinate, inver- the main stem, tributaries, lakes, ponds and tebrates such as mollusks winter in sediment ditches, hyporheic zones, riparian zones, and and other habitats, and river–lake migratory estuary. All the subsystems are interconnect- fishes stay in deep pools of the main stem. ed and interact with each other. Apparently, In spring (Figure 12B), the main stem any single/local reserve is not enough to re- floods, and river water containing large alize biodiversity conservation of the whole amounts of sediment and nutrients flows river floodplain. Therefore, it is of great im- into lakes. Hygrophytes and hydrophytes portance to conserve the huge ecosystem grow well, invertebrates revive and breed, from a holistic perspective. and migratory and resident fishes spawn and The principle of free hydrological con- then larvae migrate/drift into lakes. nectivity The principle concerns free water In summer (Figure 12C), the water level exchange among water bodies of the river in the main stem reaches a maximum. Lake- floodplain. There are four dimensions of shore is submerged and then hygrophytes hydrological connectivity (i.e., longitudinal, degenerate gradually, but hydrophytes lateral, vertical, and temporal). Free connec- flourish; fishes and invertebrates assemble tion can increase temporal-spatial heteroge- and forage in vegetation. neity of habitats, and thus species diversity, In autumn (Figure 12D), the water level and can help to prevent overgrowth of some in the main stem goes down and lake water species and thus reduce ecological disasters flows back into the river. Lakeshore emerges (e.g., blue-green algal blooms). Migratory from water and hydrophytes decline; migra- species shift habitats in different stages of life tory fishes swim to the main stem. history, and connection is of particular importance to them. To avoid habitat fragmen- Holistic Conservation of the tation, it is also necessary to maintain inter- Yangtze River Floodplain nal connection in a single water body. To implement the above principles, two Ecosystem and Strategies of scientific questions should be answered. Ecohydrological Rehabilitation First, what is the minimum protected area Conservation Principles and Key Scientific of Yangtze-connected lakes for biodiversity Questions conservation of the Yangtze River floodplain? Estimating reserve size is a key step At present, species diversity in the Yang- for effective conservation of biodiversity. At tze floodplain is continually decreasing present, only a few connected lakes maintain and water quality is continually degrading. free connection with the Yangtze main stem. Therefore, the ecosystem protection should Do they meet the requirement of biodiver- be oriented to resolve these two problems. sity conservation for the entire floodplain? Since pollution control has been discussed Second, what are the environmental flow re- extensively, we focus upon biodiversity con- quirements of the Yangtze River floodplain? status and conservation of yangtze cetaceans 23

Figure 12. Diagram of ecosystem dynamics of the Yangtze River floodplain. (From Wang and Wang 2009.)

The hydrological regimes of the floodplain Based on collected fish diversity data of waters have been altered by river damming the past 50 years, the cumulative species– and river–lake disconnection, leading to area model has been constructed (Liu and ecosystem degradation to various degrees. Wang 2010; Figure 13). Given that the total Therefore, it is urgent to understand the hy- number of fish species is 173 in this region, drological requirements of these ecosystems the minimum protected area is estimated to so as to provide the scientific basis for eco- be 14,400 km2 according to the model. Since hydrological operation of dams and sluices. the total area of existing Yangtze-connected lakes is 5,500 km2, at least 8,900 km2 of dis- Minimum Protected Area of connected lakes need to be reconnected. Yangtze-Connected Lakes Ecohydrological Operation of Dams and Species–area relationship is regarded as one Sluices of the general laws in ecology and can be used to calculate the minimum protected area. Our research shows that the existing Yang- Targeting fish conservation, we estimated tze-connected lakes do not meet the area re- the minimum protected area of the Yangtze- quirement of biodiversity conservation in the connected lakes. The reasons to choose fish Yangtze floodplain. Therefore, we must re- are (1) fish species are sensitive to hydrologi- connect enough disconnected lakes with the cal connectivity, especially migratory ones; main stem. The feasible approach is to oper- (2) fishes are K-strategists and position at ate dams and sluices in ecohydrological ways, higher trophic levels in general, and it is rea- and this should be based on the assessment of sonable to assume that most organisms can environmental flow requirements (EFRs). be conserved if fishes are under protection; In terms of macrophytes EFRs, we fo- and (3) fish data are most abundant. cused on WLFs. Macrophyte distributions 24 wang et al.

Figure 13. Estimation of minimum protected area of Yangtze-connected lakes for fish diversity conservation in the Yangtze floodplain. (From Liu and Wang 2010.) show clear zonation along elevation gradi- levels beyond the growth rate has detrimen- ents, and the life histories are closely corre- tal effects on their growth, even leading to lated with lake WLFs (Keddy and Reznicek death. The suitable submergence time is 1986). Hygrophytes are distributed at the July–August, and biomass is much higher in highest elevations and adapt to environ- areas submerged in July–August than June ments with large fluctuation amplitude. For (Zhang 2013). Submerged macrophytes pre- example, biomass of Carex in lakes with large fer lakes with small fluctuation amplitude. amplitude (e.g., Lake Dahuchi) is about 10 Underwater light is the key factor affecting times as much as those with small amplitude their growth, and the ratio of Secchi depth (e.g., Lake Yandonghu) (Zhang 2013). Hy- to mean depth should be more than 0.5 dur- grophytes germinate near lakeshore in low ing March–June in order to enable a normal water level periods, and the seedlings cannot growth (Wang et al. 2005). A certain increase tolerate submergence. During growth, hy- in water levels can promote the growth of grophytes prefer stable water levels and can- macrophytes (Zhang et al. 2013), but fast not survive if water levels rise too fast. Most rising of water levels can inhibit the growth hygrophytes mature before June, and thus or kill the plants (e.g., Keddy and Reznicek the suitable submergence time is June–July. 1986). Emergent macrophytes grow at lower eleva- With regards to fishes, we analyzed the tions than hygrophytes and are distributed at migration pattern of river-lake migratory a wide range of fluctuation amplitude. Also fishes (the four Chinese carps, i.e., Black Carp in low water level periods, they germinate Mylopharyngodon piceus, Grass Carp Cteno- in places from 1 to 7 m above lake surfaces pharyngodon idella, Silver Carp Hypophthalmi- to shallows. Taking the common reed Phrag- chthys molitrix, and Bighead Carp Hypophthal- mites australis as an example, root-sprouting michthys nobilis; Figure 14; Ru and Liu 2013). is obviously prevented at 20 cm under wa- Between late spring and early summer, the ter and absolutely fails beyond 30 cm (Cao adults breed in the main stem when the 2007). The sprouts can resist submergence water level rises sharply, and the eggs drift to a certain degree, but fast rising of water downstream and develop in the main stem. status and conservation of yangtze cetaceans 25

Figure 14. Migration diagram of four major Chinese carps in the Yangtze floodplain. (From Ru and Liu 2013.)

After spawning, adults will migrate into ing spring–summer to ensure development; floodplain lakes to feed. Some larvae and ju- and (2) open sluice gates to restore migration veniles then move passively into lakes along routes for juvenile fishes migrating into lakes with the flow. Others feed in backwaters of during April–September and for adults mi- the river and migrate later into floodplain grating into the Yangtze main stem during lakes. They grow and mature in the lakes November–December. for 3–4 years. When the water recedes in late autumn, the adults again return to the main Strategies for Holistic Conservation stem. The subadults may also leave if the Based on conservation principles and our re- lake water is not deep enough. searches, we propose the following strategies According to the above results, we pro- to conserve the Yangtze biodiversity. pose the following ways to operate dams First, establish a holistic nature reserve and sluices to meet ecohydrological require- to protect the entire Yangtze River ecosys- ments: (1) lower water levels during spring tem. In the mid to lower river floodplain, to improve germination of macrophytes, the reserve should include (1) the main stem and control rising rates of water levels dur- and connected lakes, and (2) the rehabilita- 26 wang et al. tion of at least 8,900 km2 of disconnected Acknowledgments lakes. Second, strengthen conservation actions This work is financially supported by the in the main stem and connected lakes. The State Key Laboratory of Freshwater Ecol- existing connected lakes include only two ogy and Biotechnology (2011FBZ14), the Na- large-sized lakes, Lake Poyanghu and Lake tional Natural Science Foundation of China Dongtinghu, and one medium-sized lake, (30900194 and 41371054), the Chinese Acad- Lake Shijiuhu (see Figure 1). They are biodi- emy of Sciences (KZCX1-SW-12), 973 Pro- versity hotspots. However, fishes and other grams (2008CB418006 and 2002CB412309), organisms in these lakes are facing multiple and WWF China. The map and area data of threats, such as overfishing, sand overmining, Yangtze lakes are from the Scientific Data water pollution, and unusually low water lev- Sharing Platform for Lake and Watershed els in dry seasons because of flow regulations made by the Nanjing Institute of Geography by upstream reservoirs and climate change. and Limnology, Chinese Academy of Sci- Although the Chinese central government has ences. We thank Shu-Ran Wang, Ya-Jing He, banned fishing in spring in the Yangtze main Sai-Bo Yuan, Yu Peng, Qing Yu, Xiao-Peng stem and connected lakes since 2002, the pro- Shi, and Xiao-Ke Zhang for their help. tection achieved appears to be rather limited. Numerous juvenile fish are being caught dur- References ing open fishing seasons with illegal fishing Amoros, C., and G. Bornette. 2002. Connectivity gears. We propose that commercial fishing in and biocomplexity in waterbodies of riverine the Yangtze main stem and connected lakes floodplains. Freshwater Biology 47:761–776. be banned all year round, and a core protected CAE (Chinese Academy of Engineering). 2010. area where all human activities are forbidden Stage assessment of the Three Gorges proj- should be demarcated. In addition, it is neces- ects. China Water and Power Press, Beijing. sary to reoperate the upstream dams to meet (In Chinese.) Cao, W. X. 2008. Some issues concerning protect- the ecohydrological requirements and con- ing fish resources in the Yangtze River basin. nectivity of the connected lakes. Resources and Environment in the Yangtze Third, rehabilitate disconnected lakes. basin 2:163–164. (In Chinese.) Those with good conservation potentials are Cao, W. X. 2011. The present situation and protec- of first consideration, such as oxbows,- Hon tion measures of fish resources in the Yangtze ghu and Liangzihu lakes along the middle River. Jiangxi Fishery Sciences and Technol- reaches, and Anqing lakes along the lower ogy 2: 1–4. (In Chinese.) reaches. Ecohydrological operation of dams Cao, Y. 2007. Study on impact factor and technique and sluices had better be based on EFR assess- of vegetation restoration for flood beaches wetlands along Yangtze River. Doctoral the- ments of key organisms (e.g., macrophytes, sis. Nanjing Normal University, Nanjing, Ji- benthos, fishes, water birds, and dolphins) angsu, China. (In Chinese.) and ecosystem services (e.g., clean water, food Chen, J. 2012. Evolution and water resource uti- production, and scenery). As stated above, lization of the Yangtze River. Changjiang we have done some works in this aspect, but Press, Wuhan, Hubei, China. (In Chinese.) more studies are needed. To meet the urgent Cui, L. J., Y. F. Zhai, and G. F. Wu. 2013. Dredging need of conservation at present, ecohydro- being moved southward enlarges the impact- logical operation must be implemented as ed region in Poyang Lake: the evidences from much as possible, even if knowledge is lim- multi-remote sensing images. Acta Ecologica Sinica 33:3520–3525. (In Chinese.) ited. Additionally, enclosures for aquaculture CWRC (Changjiang Water Resources Commission spread over many disconnected lakes, leading of the Ministry of Water Resources). 1999. At- to vegetation degradation and then eutrophi- las of the Changjiang River basin. SinoMaps cation. Therefore, these enclosures should be Press, Beijing. (In Chinese.) removed completely or only confined to small Du, Y., H. P. Xue, S. J. Wu, F. Ling, F. Xiao, and X. areas. H. Wei. 2011. Lake area changes in the middle status and conservation of yangtze cetaceans 27

Yangtze region of China over the 20th cen- ed lakes for fish diversity conservation in the tury. Journal of Environmental Management Yangtze River floodplain. Diversity and Dis- 92:1248–1255. tributions 16:932–940. Fang, J. Y., Z. H. Wang, Y. Li, X. W. Li, A. P. Chen, Liu, X. Q., H. Z. Wang, and X. M. Liang. 2006. C. Y. Zheng, S. Q. Zhao, S. Rao, Z. R. Tang, S. Food web of macroinvertebrate community L. Xu, H. T. Cui, H. Zeng, and F. Jiang. 2006a. in a Yangtze shallow lake: trophic basis and Ecological basis of regional biodiversity con- pathways. Hydrobiologia 571:283–295. servation of wetlands in the central Yangtze. Lytle, D. A., and N. L. Poff. 2004. Adaptation to Higher Education Press, Beijing. (In Chinese.) natural flow regimes. Trends in Ecology and Fang, M., S. H. Qi, C. X. Wu, L. S. Hu, and Q. K. Su. Evolution 19:94–100. 2006b. High resolution sedimentary record of Ma, S. S., and Q. Q. Xu. 2013. Key technical issues organochlorine pesticides in sediments core in sand excavation management in Yangtze from Honghu Lake. Geological Science and River basin. China Water Resources 10:15–18. Technology Information. 25:89–92. (In Chi- (In Chinese.) nese.) Mahoney, J. M., and S. B. Rood. 1998. Stream- Feng, W. S. 2005. Resource and environmental flow requirements for cottonwood seedling characters of a large river-joining lake, zoo- recruitment: an integrative model. Wetlands plankton. Pages 134–142 in Y. B. Cui and Z. 18:634–645. J. Li, editors. Fishery resources and conserva- MWR (Ministry of Water Resources of China). tion of environment in lakes of the Changji- 2004–2013. China water resources bulletin ang River basin. Science Press, Beijing. (In (2003–2012). China Water and Power Press, Chinese.) Beijing. (In Chinese.) IGSNRRCAS (Institute of Geographic Sciences OECD (Organization for Economic Co-operation and Natural Resources Research, Chinese and Development). 2006. Eutrophication of Academy of Sciences), Changjiang Institute waters: monitoring, assessment and control. of Water Resources and Hydropower Re- Prepared by Soil & Water Conservation Soci- search), and Changjiang Waterway Plan- ety of Metro Halifax. Available: http://lakes. ning Design and Research Institute. 1985. chebucto.org/TPMODELS/OECD/oecd. The mid–lower reaches of the Yangtze River html. (May 2015). channel characteristics and their evolution. Pan, B. Z., H. J. Wang, X. M. Liang, and H. Z. Science Press, Beijing. (In Chinese.) Wang. 2009. Phytoplankton chlorophyll a in Junk, W. J. 2005. Flood pulsing and the linkages the Yangtze-connected lakes: present state, between terrestrial, aquatic, and wetland sys- influencing factors and empirical relation- tems. Internationale Vereinigung fur Theo- ships. Fresenius Environmental Bulletin retische und Angewandte Limnologie Ver- 18:1894–1900. handlungen 29:11–17. Pan, B. Z., H. J. Wang, X. M. Liang, and H. Z. Keddy, P. A., and A. A. Reznicek. 1986. Great lake Wang. 2011a. Macrozoobenthos in Yangtze vegetation dynamics: the effects of fluctuat- floodplain lakes: patterns of density,- bio ing water levels and buried seeds. Journal of mass, and production in relation to river con- Great Lakes Research 12:25–36. nectivity. Journal of the North American Ben- Li, J., Z. Q. Xia, H. C. Dai, W. Yi. 2013. Effect of thological Society 30:589–602. the Three Gorges Reservoir initial filling on Pan, B. Z., H. J. Wang, H. Z. Wang, and Z. Y. Wang. downstream habitat suitability of the typi- 2011b. Macrozoobenthic assemblages in rela- cal fishes. Journal of Hydraulic Engineering tion to environments of the Yangtze-isolated 44:892–900. (In Chinese.) lakes. Frontiers of Environmental Science and Lin, Z. X. 2003. Analysis of water environmental Engineering 6:246–254. change in Taihu watershed. Journal of Lake Pan, B. Z., H. Z. Wang, and H. J. Wang. 2014. A Sciences 14:97–103. (In Chinese.) floodplain-scale lake classification based on Liu, X. Q., and H. Z. Wang. 2008. Food web of ben- characteristics of macroinvertebrate assem- thic macroinvertebrates in a large Yangtze blages and corresponding environmental River-connected lake: the role of flood distur- properties. Limnologica 49:10–17. bance. Fundamental and Applied Limnology Poff, N. L., and D. D. Hart. 2002. How dams vary 171:297–309. and why it matters for the emerging science Liu, X. Q., and H. Z. Wang. 2010. Estimation of of dam removal an ecological classification minimum area requirement of river-connect- of dams is needed to characterize how the 28 wang et al.

tremendous variation in the size, operational Lakeshore overview of Lake Chaohu and eco- mode, age, and number of dams in a river logical rehabilitation schemes for shoreline basin influences the potential for restoring and littoral zones. Resources and Environ- regulated rivers via dam removal. BioScience ment in the Yangtze basin 21(Z2):62–68. (In 52:659–668. Chinese.) Rao, Q. Z., editor. 1956. Basic knowledge of lake Wang, H. Z., and H. J. Wang. 2009. Ecological ef- survey. Science Press, Beijing. (In Chinese.) fects of river-lake disconnection and restora- Rood, S. B., G. M. Samuelson, J. H. Braatne, C. R. tion strategies in the mid–lower Yangtze ba- Gourley, F. M. Hughes, and J. M. Mahoney. sin. Pages 379–396 in Z. Y. Wang, D. G. Shao, 2005. Managing river flows to restore flood- X. J. Shao, Y. J. Lu, and X. H. Duan, editors. plain forests. Frontiers in Ecology and the En- Integrated management of water, sediment vironment 3:193–201. and ecology in the Yangtze River basin. Sci- Ru, H. J., and X. Q. Liu. 2013. River-lake migra- ence Press, Beijing. (In Chinese.) tion of fishes in the Dongting Lake area of the Wang H. Z., Z. C. Xie, X. P. Wu, and Y. L. Liang. Yangtze floodplain. Journal of Applied Ich- 1999. A preliminary study of zoobenthos thyology 29:594–601. in the Poyang Lake, the largest freshwater Shen, Z. Y., J. F. Niu, J. Qi, L. L. Wang, Z. W. Tang, lake of China, and its adjoining reaches of and J. L. Feng, editors. 2008. Characteristics Changjiang River. Acta Hydrobiologica Si- and ecological risks of water pollution in nica 23(Supplement):132–138. (In Chinese.) typical segments of the mid-Yangtze River. Wang, H. Z., H. J. Wang, X. M. Liang, and Y. D. China Environmental Science Press, Beijing. Cui. 2006. Stocking models of Chinese mit- (In Chinese.) ten crab (Eriocheir japonica sinensis) in Yangtze Shi, Y., D. W. Yang, L. Ning, and J. A. Ma. 2011. lakes. Aquaculture 255:456–465. Assessment of effects of the operation of Wang, H. Z., H. J. Wang, X. M. Liang, L. Y. Ni, X. Three Gorges Dam on the flood control in Q. Liu, and Y. D. Cui. 2005. Empirical mod- the mid–lower Changjiang River. Changjiang elling of submersed macrophytes in Yangtze Press, Wuhan, Hubei, China. (In Chinese.) lakes. Ecological Modelling 188:483–491. Shu, F. Y., H. J. Wang, L. X. Liu, and H. Z. Wang, Wang, H. Z., Q. Q. Xu, Y. D. Cui, and Y. L. Li- 2009. Assessment of species status of Mol- ang. 2007. Macrozoobenthic community of lusca in the mid–lower Yangtze lakes. Acta Poyang Lake, the largest freshwater lake of Hydrobiologia Sinica 33:84–91. (In Chinese.) China, in the Yangtze floodplain. Limnology STSGDFCR (Survey Team of Spawning Grounds 8:65–71. of Domestic Fishes in Changjiang River). Wang, J., and H. Y. Chen. 2010. Water regime vari- 1982. A survey on the spawning grounds ation of the middle and low reaches of Yang- of the “four famous Chinese carps” in the tze River during the water storage period of Changjiang River after dammed by the key Three Gorges Reservoir and its countermea- water control project at Gezhouba. Journal of sures. China Water Resources 19:15–17. (In Fisheries of China 6:287–304. (In Chinese.) Chinese.) Wang, H. J. 2007. Predictive limnological re- Wang, J. T. 1959. The embankment project of the searches on small- to medium-sized lakes Yangtze River. Yangtze River 5:61–64. along the mid–lower Yangtze River. Doctoral Wang, S. M., and H. S. Dou. 1998. Lakes of China. dissertation. University of Chinese Academy Science Press, Beijing. (In Chinese.) of Sciences, Beijing. (In Chinese.) Ward, J. V. 1997. An expansive perspective of riv- Wang, H. J., X. M. Liang, P. H. Jiang, J. Wang, S. erine landscapes: pattern and process across K. Wu, and H. Z. Wang. 2008. TN:TP ratio scales. GAIA - Ecological Perspectives for Sci- and planktivorous fish do not affect nutrient- ence and Society 6:52–60. chlorophyll relationships in shallow lakes. Ward, J. V., Malard, F., and K. Tockner. 2002. Freshwater Biology 53:935–944. Landscape ecology: a framework for integrat- Wang, H. J., H. Z. Wang, X. M. Liang, and S. K. ing pattern and process in river corridors. Wu. 2014. Total phosphorus thresholds for Landscape Ecology 17:35–45. regime shifts are nearly equal in subtropi- Wei, X., Y. Du, S. Cai, T. Zhang, and T. Liu. 2007. cal and temperate shallow lakes with mod- On the evolution of lakes in Hubei Province erate depths and areas. Freshwater Biology and countermeasures. Journal of Lake Sci- 59:1659–1671. ences 19:530–536. (In Chinese.) Wang, H. Z., C. L. Song, X. Q. Liu, and K. Li. 2012. Wu, G. F., and L. J. Cui. 2008. Remote sense-based status and conservation of yangtze cetaceans 29

analysis of the sand dredging impact on wa- Zhang, D. B., S. P. He, G. X. Hu, and G. P. Hu. ter clarity in Poyang Lake. Acta Ecologica Si- 2013. Preliminary analysis on runoff variation nica 28:6113–6120. (In Chinese.) of middle Yangtze River after impoundment Wu, H. W., and C. C. Jao. 1958. Fish farming in the of Three Gorges Reservoir. Yangtze River 1: shallow lakes in China. Verhandlungen des 1–3. (In Chinese.) Internationalen Verein Limnologie 13:765– Zhang, G., L. Wu, H. T. Li, M. Liu, F. Cheng, B. R. 769. Murphy, and S. G. Xie. 2011. Preliminary evi- Wu, Z. G., and J. Li. 2010. Seeking truth and be- dence of delayed spawning and suppressed ing pragmatic for reform and innovation to larval growth and condition of the major develop a new situation of sand mining man- carps in the Yangtze River below the Three agement of Yangtze River. Yangtze River Gorges Dam. Environmental Biology of Fish- 41:31–36. (In Chinese.) es 93:439–447. Xie, P. 2009. Reading about the histories of cyano- Zhang, X. K. 2013. Water level fluctuation require- bacteria, eutrophication and geological evo- ments of plants in the Yangtze floodplain lution in Lake Chaohu. Science Press, Beijing. lakes. Doctoral dissertation, University of (In Chinese.) Chinese Academy of Sciences, Beijing. (In Xie, S. G., Z. J. Li, J. S. Liu, S. Q. Xie, H. Z. Wang, Chinese.) and B. R. Murphy. 2007. Fisheries of the Zhang, X. K., X. Q. Liu, and H. Z. Wang. 2014. De- Yangtze River show immediate impacts of veloping water level regulation strategies for the Three Gorges Dam. Fisheries 32:343–344. macrophytes restoration of a large Yangtze- Xu, J. J., and J. Chen. 2013. Study on the impact of disconnected lake, China. Ecological Engi- Three Gorges Reservoir on Poyang Lake and neering 68:25–31. some proposals. Journal of Hydraulic Engi- Zhang, Z. L., and L. Z. Huang. 2008. Influence of neering 7:3. (In Chinese.) quarrying in Poyang Lake in the ecological Xu, Q. Q., H. Z. Wang, and S. P. Zhang. 2003. The environment. Jiangxi Hydraulic Science and impact of overstocking of mitten crab, Erio- Technology 34:7–10. (In Chinese.) cheir sinensis, on lacustrine zoobenthic com- Zhao, J., and C. K Chen, editors. 1999. China’s ge- munity. Acta Hydrobiologica Sinica 27:41–46. ography. Higher Education Press, Beijing. (In (In Chinese.) Chinese.) Yang, X. D., J. Shen, X. H. Dong, and E. F. Liu. Zhao, W. H. 2010. Macroecological patterns of 2006. Nutrition state evolution and lake eco- macrozoobenthos in rivers of China and envi- logical response of shallow lakes in history ronmental flow requirements in lower reach- period in mid–lower Yangtze River—taking es of the Yellow River. Doctoral dissertation. Lake Long’gan and Lake Tai’bai for example. University of Chinese Academy of Sciences, Science in China Series D: Earth Sciences Beijing. (In Chinese.) 35(Supplement II):45–54. (In Chinese.) Zheng, H., P. D. Clift, P. Wang, R. Tada, J. Jia, M. Yi, B. L., Z. T. Yu, and Z. S. Liang. 1988. Gezhouba He, and F. Jourdan. 2013. Pre-Miocene birth water control project and four famous fish in of the Yangtze River. Proceedings of the Na- Yangtze River. Hubei Science and Technolo- tional Academy of Sciences of the United gy Press, Wuhan, Hubei, China. (In Chinese.) States of America 110:7556–7561. Yu, W. C., and J. Y. Lu. 2005. The evolution and Zhu, K. Z. 1972. A preliminary study of the cli- regulation of the Yangtze River. China Water matic changes in China since recent 5000 and Power Press, Beijing. (In Chinese.) years. Acta Archaeological Sinica 1:15–38. (In Yu, X. D., T. H. Luo, and H. Z. Zhou. 2005. Large- Chinese.) scale patterns in species diversity of fishes in Zhu, M. 1996. Variation trend and protection steps the Yangtze River basin. Biodiversity Science of water quality in Taihu Lake. Journal of 13:473–495. (In Chinese.) Lake Sciences 8:133–138. (In Chinese.)