Stream Networks and Knickpoints in the Sanjiangyuan Region

10.2478/v10060-008-0067-5

Annals of Warsaw University of Life Sciences – SGGW Land Reclamation No 42 (1), 2010: 79–91 (Ann. Warsaw Univ. of Life Sci. – SGGW, Land Reclam. 42 (1), 2010)

Stream networks and knickpoints in the Sanjiangyuan Region

ZHAOYIN WANG1, GUO-AN YU2, GARY BRIERLEY3, LE LIU4 1State Key Laboratory of Hydroscience and Engineering, Tsinghua University, China 2Institute of Geographical Sciences and Natural Resources Research, Chinese Academy of Sciences 3School of Environment, University of Auckland 4Department of Hydraulic Engineering, Tsinghua University, China

Abstract: Stream networks and knickpoints in INTRODUCTION the Sanjiangyuan Region. The source area of the Lancang (Mekong), Yangtze and Yellow rivers is The 6,300 km Yangtze River and 5,464 km named in Chinese Sanjiangyuan (source of three Yellow River are the longest and second rivers). Geographical characteristics of these rivers, and various rivers at the margin of the Qing- longest rivers in China. The Lancang hai-Tibetan Plateau (Jialing, Minjiang, Dadu, River (Mekong River) is the largest Yalong and Jinsha) are summarized from fi eld international river in Asia. The source investigations along with digital elevation model area of the three rivers is named in (DEM) analyses and satellite images. Uplift of Chinese Sanjiangyuan (source of three the Qinghai-Tibetan Plateau has resulted in an rivers). It extends on the Qinghai-Tibet asymmetrical distribution of tributaries for stream o o networks at the margins of the plateau. Almost all plateau from N31 39’ to N36 16’ and o o tributaries join the trunk stream from the north- from E89 24’ to E102 23’. Altitude varies west. Three types of drainage network are evi- from 2,000-5,000 m, with a mean altitude dent: plume, nervation and dendritic. In general, of about 4,000 m. The total area of San- plume networks have a large number of fi rst order jiangyuan is 360,000 km2. About 49% streams. Nervation networks have a main stream of Yellow River water, 15% of Lancang and parallel tributaries. Dentritic networks are characrterized by continuous bifurcation and have River water and 2.5% of the Yangtze a branch-like appearance. Most stream networks River water come from this region in the Sanjiangyuan region are of the nervation (http://www.snowland-great-rivers. type. Several large knickpoints are evident along org/threerivers.htm). There are more the longitudinal profi les of the Yellow and Yang- than 180 streams, 16,500 lakes, a total tze Rivers. Channel aggradation immediately up- of 73,300 km2 of swamp, and numerous stream of these knickpoints marks a transition in river processes from vertical bed evolution (i.e. snow mountains and glaciers with a total 2 incision) to horizontal channel adjustment (and area of 2,400 km in the region. associated braided and anabranching channels). Northward movement of the Indian Plate, and its collision with the Eurasian Key words: Qinghai-Tibetan Plateau, knickpoint, Plate, has uplifted the Qinghai-Tibetan drainage network, Sanjiangyuan, step-pool. Plateau, which is referred to as the ‘third 80 Zhaoyin Wang et al. pole’ or the ‘roof of the world’. Uplift of ciated high sediment transport rates, has the Qinghai-Tibetan Plateau has created resulted in various ecological problems drier and colder conditions (Tang and in the Sanjiangyuan region. Dong 1997). It has also brought about This paper overviews summary results large-scale and profound environmental from fi eld and DEM-based analyses on changes in surrounding areas, including stream network patterns in the Sanjiangy- the formation of deserts in northwestern uan region and knickpoint processes at China (Dong et al. 1994). the margin of the Qinghai-Tibetan Pla- Various stream network patterns have teau. Results from these scientifi c inves- formed under different environmental tigations will hopefully provide effective conditions in the region. Rather than guidance for integrated management of developing a typical dendritic (tree-like) natural resources in the Sanjiangyuan form, stream networks have a nervation region, aiding sustainable development (vein-like) form (Wang et al. 2009b). programmes that strive to reconcile eco- Hortonian laws of stream numbers, logical protection with sustainable use of lengths and drainage areas vary mark- land and water resources. edly across the region, with notable dif- ferences in glaciated and non-glaciated METHODS areas. Indeed, glacial and fl uvial erosion are the main driving forces in the deve- Geographical and geometrical characte- lopment of complex stream-lake network ristics were measured in the fi eld using in the region. Global Positioning System (GPS) and Deformation of landscapes at the pla- laser range meter. A 1995 digital map teau margin has resulted in steep, highly of China with a scale ratio of 1:250.000, dynamic rivers, many of which are cha- 1960s topographic maps with a scale ratio of 1:50 000, and 2004 digital topographic racterized by typical step-pool system SRTM-3 data were used to construct sequences (Chin 2002; Costa and Schus- longitudinal profi les and cross sections ter 1988; Korup 2004; Wang et al. 2009a). of various rivers. This was completed An approximate topographic equilib- using ARCGIS. GoogleEarth was used to rium has been maintained as rapid fl u- provide global satellite images at varying vial incision into bedrock and headward resolutions. In this study, the “eye alti- migration of knickpoints strives to keep tude” was fi xed to ensure that all data were balance with tectonic uplift (Korup et al. collected at the same resolution. 2006; 2009). This has further accentua- ted slope instability, so landslides and debris fl ows are common (Ouimet et al. ASYMMETRICAL DISTRIBUTION 2007). High sediment loads carried by OF TRIBUTARIES the Yellow, Yangtze and Lancang Rivers According to the British Geological refl ect high rates of landscape denuda- Survey, the Indian Plate moves northward tion. This has increased sediment trans- at a rate of 5 cm/year (Chen and Gavin, port by rivers and offshore deposition. 2008). Its collision with the Eurasian Plate Landscape instability brought about by has resulted in the uplift of the Himalaya incision and mass movements, and asso- and the Qinghai-Tibetan Plateau, and Stream networks and knickpoints... 81 associated earthquake activity. The rate Jialing) are all tributaries of the Yangtze of horizontal movement is about 5 cm/yr River. The rising plateau has increased at Himalaya, 4 cm/yr in the source area the gradient of these rivers, inducing of the Lancang and Yangtze Rivers, to incision that often extends over 2 km about 2 cm/yr at the Qilian Mountain. deep. Uplift has caused dramatic chan- As a result the Himalaya rises at a rate ges to stream networks. All major rivers of 21 mm/yr while the Qilian Mountains in this region fl ow from northwest to rise at a rate of 5 mm/yr. The Sichuan southeast, refl ecting the uplift of the pla- basin, located to the east of the plateau, teau and relatively stationary position of is stationary. This pronounced variability the Sichuan Basin to the east. in the rates of horizontal movement has Asymmetry of stream networks is also resulted in many active faults, such as evident in small watersheds. Figure 2 the Kalakunlun Fault to the west, Kunlun shows the stream network of the Xihan- fault in the center, Altyn fault to the north, shui River (the upstream area of the Jial- and Xianshuihe and Longmenshan faults ing River) on the northeastern margin of to the east. The Wenchuan earthquake the plateau. Once more, all large tribu- on May 12, 2008 occurred along the taries join the river from the northwe- Longmenshan fault. stern side, creating a very asymmetrical Streams along the east margin of the stream network. More than 90% of the Qinghai-Tibetan Plateau and the Long- drainage area is on the western side of menshan Fault are shown on Figure 1. the river. This situation is repeated for The four major rivers of Sichuan Pro- many stream networks at the margins of vince (Minjiang, Tuojiang, Fujiang and the Qinghai-Tibetan Plateau.

FIGURE 1. Streams on the east margin of the Qinghai-Tibetan Plateau and the Longmenshan Fault 82 Zhaoyin Wang et al.

FIGURE 2. Asymmetrical distribution of tributaries and drainage area of the Xihanshui River

FIGURE 3. Wenchuan earthquake changed the bed slope and cut off the ﬂ ow in a tributary stream of the Shenxi Ravine in the Minjiang River basin

Shenxi Ravine, a tributary stream in east (Fig. 3). The most recent uplift phase the Minjiang River basin, is located along of the plateau, the Wenchuan earthquake, the Longmenshan Fault which separates induced a 4 m vertical displacement along the Qinghai-Tibetan Plateau to the north- this fault, as the area to the northwest was west and the Sichuan basin to the south- uplifted. It triggered numerous landslides Stream networks and knickpoints... 83 and avalanches (Wang et al. 2009a). orders). However, various researchers Prior to the earthquake, the stream shown have questioned the universality of Hor- in Figure 3 was a perennial stream that ton’s law. For instance, Kirchner (1993) fl owed into the Shenxi Ravine. Tectonic argued that Horton’s ratio may vary for motion cut off surface runoff and changed different stream networks. Also, Liu and the slope of the stream. Such episodes Wang (2008) found that Horton’s ratio have been repeated many times during the is constant only for the stream orders millions of years over which uplift of the higher than 8. For lower stream orders, Qinghai-Tibetan Plateau has taken place. Horton’s ratio varies with stream order This has resulted in very asymmetrical and is very different for different river stream networks. networks. For example, a large number of fi rst order streams fl ow into a second STREAM NETWORKS ON THE order stream in a plume network (Fig. QINGHAI-TIBETAN PLATEAU 4a). In contrast, a nervation network has a main stream and parallel tributa- Horton (1945) pioneered quantitative ries (Fig. 4b). Finally, dendritic networks studies of river morphology by introdu- have continuous bifurcation and look cing his laws of stream number, Nm, i.e. like tree branches (Fig. 4c); these are the notionally ‘typical’ networks to which BZ (1) NAeZ Horton’s Law applies. Most stream networks in the San- where ω is stream order; NZ is the number of ω-th order streams; and A and B are jiangyuan region are nervation, with constants. This empirical law was not some plume networks. For example, derived from basic theories in physics Figure 5a shows a plume stream network or other fundamental theories. However, in the source of the Dadu River, a tribu- although arguments about the Horton’s tary of the Yangtze River. Figures 5b and Laws persist, the validity of equations c show nervation networks in the source has been independently verifi ed and of the Yellow River near Kesheng Town they have been accepted as basic laws in and Maqin County. river morphology (Ciccacci et al. 1992; Liu and Wang (2008) calculated Hor- Kinner and Moody 2005). ton’s ratios for typical plume and nervation networks in the loess plateau area of The bifurcation ratio, RB, characteri- zes the ratio of the number of streams of China and typical dendritic networks in differing stream orders according to the northeast China (Fig. 6). The bifurcation following equation: ratio for plume networks is more than 10 for stream orders less than 3. This is much

NZ B greater than the value for dendritic net- ReB (2) N works (about 4) and nervation networks Z1 (about 5). For stream orders higher than

According to Horton’s Law, RB is 4, the bifurcation ratio reduces and con- invariant with stream orders and net- verges for all three types of networks, works (i.e. Horton’s ratio is the same for and approaches 4 for 8th order streams. different rivers and for different stream The ratio for dendritic networks is almost 84 Zhaoyin Wang et al. a b c

FIGURE 4. a – Plume network in the source region of the Dadu River; b – Nervation network near Dari County; c – Typical dendritic network in the Songhua River basin in northeastern China

a b c

FIGURE 5. Stream networks extracted from satellite images: a – source of Dadu River; b – source of Yellow River near Kesheng (Maqin County); and c – source of Yellow River near Maqin county town invariant for different stream orders but network of the Dadu River shown in it has a great range for plume networks. Figure 5a is consistent with results from Hence, while dendritic networks may plume networks, while networks in the show remarkable similarity in physical source region of the Yellow River near structure between large and small water- Kesheng and Maqin are similar to results sheds, these relationships may be very from nervation networks (these three different for plume networks. The stream networks are plotted on Fig. 6). Stream networks and knickpoints... 85

15 B R 10

0 0123456789 ȍ O Plume NervationNervation Dentritic TributaryTributary of of the the YR YR near near Maqin Maqin Tributary of of the the YR YR near near Kesheng Kesheng SourceSource of of the the Dadu Dadu River River FIGURE 6. Comparison of bifurcation ratios for differing types of stream network, highlighting com- parisons for the networks shown in Figure 5

It would be interesting to relate these is more than 2000 m. Incision has resulted results of stream network analyses to in very steep and unstable slopes adjacent differing lithological considerations and to channels. Earthquakes and rainstorms associated linkages to tectonic setting. trigger many landslides and avalanches, Indeed, comparison of networks atop some of which may dam rivers and form the Qinghai-Tibetan Plateau with those barrier lakes. These natural dams protect in the Sichuan Basin could be used to bedrock from fl uvial incision and knick- assess tectonic imprints upon stream net- point migration, helping to stabilize the works, prospectively determining vari- margin of the plateau in concert with the ous stages in the evolution of drainage effects of localized rock uplift (Korup et networks and patterns. al. 2009). The slope below a landslide dam is steep but the upstream slope is very KNICKPOINTS AND THEIR EFFECT gentle. Hence, these knickpoints prohibit ON RIVER STABILITY a river from eroding its bed and incising upstream. This continues as long as Uplift of the Qinghai-Tibetan Plateau the landslide dams are not broken and has caused signifi cant bed incision along scoured away. A landslide dam may de- almost all rivers. Incision is deepest at the velop into a knickpoint if it is stabilized margin of the plateau. The longitudinal under the long-term action of fl ow. Large profi le of the Minjiang River is related boulders and cobbles are rearranged to to the elevation of the adjacent mountain form step-pool systems (Todd and Mike range in Figure 7. Depth of incision 2003; Whittaker and Jaeggi 1982). These shown by the difference in elevation be- features create great resistance and con- tween the mountains and the channel bed sume considerable fl ow energy, thereby 86 Zhaoyin Wang et al. protecting the channel bed from erosion ries of these features, may stabilize large as well as creating habitat for benthic knickpoints as several steps with very invertebrates (Milzow et al. 2006; Nicko- high slope (Wang et al. 2009d). The total lotsky and Pavlowsky 2007; Wang et al. water head of a large knickpoint can be 2009c). as large as several hundred meters. The Large knickpoints at the margin of the Hutiaoxia (Tiger Leaping Gorge) land- Qinghai-Tibetan Plateau range in eleva- slide and avalanche dam on the Jinsha tion from 1000 to 4000 m. Preservation River has a water head of 213 m. Several of landslide and avalanche dams, or a se- large knickpoints at Hutiaoxia, Deqin,

4500 4000 River bed Songpan 3500 Mountain 3000 Avalanche Diexi 2500 Wenchuan 2000 Dujiangyan Elevation (m) Elevation 1500 1000 500 0 0 100 200 300 400 500 600 700 800 Distance to the Yangtze River (km)

FIGURE 7. Longitudinal bed proﬁ le and elevation of the adjacent mountain ranges along the Minjiang River

2600 Mangkang

2400 Deqin 2200

2000 Hutiaoxia

Elevation (m) 1800 Muli 1600

1400 350 450 550 650 750 850 950 Distance (km) FIGURE 8. Bed profi le of the Jinsha River showing the large knickpoints at Hutiaoxia (Leaping Tiger Gorge), Deqin and Mangkang Stream networks and knickpoints... 87 and Mangkang are shown along the lon- immediately downstream of Dari County gitudinal profi le of the Jinsha River in town and downstream of Kesheng Town. Figure 8. The Hutiaoxia landslide and Several landslide dams, with large bar- avalanche dam created a large barrier rier lakes upstream, characterize these lake upstream (Fig. 9). Although the dis- knickpoints. These barrier lakes have charge of the river is about 4,000 m3/s the been infi lled by this sediment-laden river. water surface in the barrier lake was as Figure 10b shows a part of the Kesheng smooth as a mirror because of the huge knickpoint, which consists of several depth and low fl ow velocity. The lake landslide dams and barrier lakes. The traps all bedload and a large portion of transition from landslide dams into large the suspended load. The lake has not yet knickpoints has greatly infl uenced river been fi lled because it was created quite profi les and landscape evolution. The recently and the relatively low sediment longitudinal bed profi le of the Yellow loads from upstream. A resilient step-pool River in the 1,400 km long reach is very system with huge hydraulic jumps has unusual, with a convex upward curve developed downstream of the barrier lake, caused by the two large knickpoints. consuming most of the fl ow energy and Harkins et al. (2007) report a similar protecting the river bed from erosion. phenomenon for tributaries of the Yellow Large knickpoints may totally change River. Downstream reaches are steep but patterns of fl uvial processes and river incision is controlled by step-pool sys- morphology. Figure 10a shows the bed tems. Upstream reaches have adjusted to profi le of the Yellow River from the a new, unchanging base level. source (Erlin Lake) to Longyangxia Cross sections from the upstream sec- Dam – the uppermost large dam on the tion and the landslide dam section of the river. Two large knickpoints are evident: Dari and Kesheng knickpoints are pre- a b

FIGURE 9. a – Hutiaoxia barrier lake; b – Part of the step-pool system on the Hutiaoxia landslide dam 88 Zhaoyin Wang et al. a b 4300 Eling Lake 3000

4000 Dari

3700 Kesheng 2850 3400 Elevation (m) Elevation

3100 (m) Elevation 2700

2800 Longyangxia Dam Longyangxia Dam 2500 2550 0 200 400 600 800 1000 1200 1400 1200 1300 1400 Distance (km) Distance (km) FIGURE 10. a – Bed proﬁ le of the Yellow River from the source (Erlin Lake) to Longyangxia Dam; b – Part of the Kesheng knickpoint consisting of several landslide dams and barrier lakes a b 4500 4400 4100 4300 4050 4200 4000 4100

3950 Elevation (m) 4000

Elevation (m) Elevation 3900 3900 3850 3800 0 800 1600 2400 3200 4000 4800 5600 6400 0 1000 2000 3000 4000 5000 Distance (m) Distance (m) c d 4100 4000 3700 3900 3600 3800 3700 3500 3600 Elevation (m) Elevation

3500 (m) Elevation 3400 3400

3300 3300 0 2000 4000 6000 8000 10000 12000 0 2000 4000 6000 8000 10000 Distance (m) Distance (m) FIGURE 11. a – Cross section upstream section of the Dari knickpoint; b – Cross section of a landslide dam at the Dari knickpoint; c – Cross section upstream of the Kesheng knickpoint; d – Cross section of a landslide dam of the Kesheng knickpoint Stream networks and knickpoints... 89 sented in Figure 11. Landslides evident CONCLUSIONS on the left hand side of Figure 11b and the right hand side of Figure 11d likely Over time the uplift of the Qinghai-Ti- dammed the river and infl uenced the ge- betan Plateau and the relative movement neration of knickpoints. Accumulation of of the Earth’s crust at the margins of the huge boulders on the bed formed a stable plateau has modifi ed river alignment, re- step-pool system that is strong enough to sulting in asymmetrical stream networks resist the fl ow and protect the bed from for the Minjiang, Tuojiang, Fujiang erosion. Long term aggradation following and Jialing Rivers and their tributaries. knickpoint formation has resulted in wide Almost all of these rivers fl ow from the and shallow cross sections in upstream northwest to the southeast. Asymmetry areas, whereas the landslide dam section is also evident for midsize rivers (e.g. itself is narrow and deep. The distribution more than 90% of the drainage area of of channel processes and resulting river the Xihanshui River is on the western morphologies, have adjusted to these side of the river). damming events. Localized aggradation Much of the Sanjiangyuan region is upstream of knickpoints has brought dominated by nervation stream networks, about a transition from vertical bed with some plume types. The bifurcation evolution to horizontal fl uvial process. ratio of these networks is high for second Reduction in slope and accumulation of order and third order streams and reduces fi ne-grained sediments has facilitated the with increasing in stream order. The ratio development of anabranching and brai- is around 4 for stream order large than 8. ded channel planforms (e.g. just upstream Uplift of the Qinghai-Tibetan plateau of Dari, Fig. 12). has caused extensive channel bed inci-

FIGURE 12. Satellite image of the branching channels upstream of the Dari knickpoint 90 Zhaoyin Wang et al. sion along almost all rivers. In some CICCACCI S., D’ALESSANDRO L., FREDI instances this exceeds 2,000 m. This has P., 1992: Relation between morphometric cha- created great potential for landslides and racteristics and denudation processes in some drainage basins of Italy. Zeitschrift für Geomor- avalanche activity. A landslide dam may phologie. 36(1): 53–67. develop into a knickpoint if it is stabilized COSTA J.E., SCHUSTER R.L., 1988: The for- by the long-term action of fl ow. Large mation and failure of natural dams. Geological knickpoints have totally changed the dis- Society of America Bulletin. 100: 1054–1068. tribution of fl uvial processes and resul- DONG GUANGRONG, WANG GUIRONG, CHEN HUIZHONG, YAN MANCUN, JIN ting river morphology in some reaches JIONG, WANG YUE, 1994: The formation of the source zone of the Yellow River. and evolution of the deserts in China and their Longitudinal profi les are reshaped by relation to the uplifting of Qinghai-Tibet Pla- the distribution of knickpoints. Although teau. Proceedings of Qinghai-Tibetan Plateau reaches downstream of knickpoints are and Global Variations, 13–29. (in Chinese) HARKINS N., KIRBY E., HEIMSATH A., ROB- steep, incision is controlled by step-pool INSON R., REISER U., 2007: Transient fl uvial sequences that consumes the majority incision in the headwaters of the Yellow River, of the fl ow energy and protect the river northeastern Tibet, China. Journal of Geo- bed from erosion. Upstream reaches physical Research. F. Earth Surface. 112, F3. have a new, unchanging base level. This HORTON R.E., 1945: Erosional development brings about a transition from degrada- of streams and their drainage basins: hydro- physical approach to quantitative morphology. tion to aggradation and from vertical Geological Society of America Bulletin. 56: bed evolution to horizontal fl uvial pro- 275–370. cess. Multiple and unstable channels are KINNER D.A., MOODY J.A., 2005: Drainage prominent in the reaches upstream of the networks after wildfi re. International Journal knickpoints. of Sediment Research. 20(3): 194–201. KIRCHNER J.W., 1993: Statistical inevitability of Horton’s laws and the apparent random- Acknowledgement: The study is support- ness of stream channel networks. Geology, 21: ed by the Sino-New Zealand collaboration 591–594. research project, Ministry of Science and KORUP O., MONTGOMERY, D.R., HEWITT Technology of China (2008CB425803 K., 2009: Spatial clustering of natural dams at and 2007SHZ0901034). We are grateful the Tibetan Plateau margin in rivers draining the Himalayan syntaxes. Geophysical Research to Prof. Heqing Huang, Chinese Acade- Abstracts. 11: EGU2009-8344 my of Sciences, for his support to this KORUP O., STROM A., WEIDINGER J., 2006: study. Fluvial response to large rock-slope failures: Examples from the Himalayas, the Tien shan, REFERENCES and the southern Alps in New Zealand. Geo- morphology, 78(1-2): 3–21. CHEN J., GAVIN H., 2008: Finite Fault Model KORUP O., 2004: Geomorphometric characteris- of the May 12, 2008 Mw 7.9 Eastern Sichuan, tics of New Zealand landslide dams. Engineer- China Earthquake. United States Geological ing Geology. 73: 13–35. Survey – National Earthquake Information LIU HUAIXIANG. Research on the Distribution Center. http://earthquake.usgs.gov/, Retrieved of Streambed Structures and Their infl uence on 2008-05-15. Fluvial Morphology. Doctoral dissertation, CHIN A., 2002: The periodic nature of step-pool Tsinghua University, 2009 (in Chinese). mountain streams. American Journal of Sci- LIU H.X., WANG Z.Y., 2008: Morphological fea- ence. 302: 144–167. ture and distribution of typical river networks. Stream networks and knickpoints... 91

Journal of Hydraulic Engineering. 38(11): WANG ZHAO-YIN, WAI ONYX W., CUI 1354–1357 (in Chinese). PENG, 1999: Field investigation into debris MILZOW C., MOLNAR P., McARDELL B.W., fl ow mechanism. International Journal of BURLANDO P., 2006: Spatial organization Sediment Research. 4: 10–22. in the step-pool structure of a steep mountain WHITTAKER J.G., JAEGGI N.R., 1982: Origin stream (Vogelbach, Switzerland). Water Resour- of step-pool system in mountain streams. Jour- ces Research. 42: W04418. nal of the Hydraulics Division, ASCE. 108: NICKOLOTSKY A., PAVLOWSKY R.T., 2007: 758–773. Morphology of step-pools in a wilderness headwater stream: The importance of standar- Streszczenie: Rozwój progów skalnych oraz mor- dizing geomorphic measurements. Geomor- fologia rzek w regione Sanjiangyuan. W artykule phology. 83: 294–306. przedstawiono proces rozwoju progów skalnych OUIMET W.B., WHIPPLE K.X., ROYDEN, L.H. wzdłuż rzek Yellow i Yangtze. Wykazano wpływ SUN ZHIMING, CHEN ZHILIANG, 2007: progów skalnych na zmiany morfologiczne The infl uence of large landslides on river inci- dna koryt rzek. Głównie skoncentrowano się na sion in a transient landscape: Eastern margin of zmianach agradacyjnych powyżej omawianych the Tibetan Plateau (Sichuan, China). Geologi- progów. Progi te są wyznacznikami zmian cha- cal Society of America Bulletin. 119 (11/12): rakteru koryta rzecznego, z pionowo wciętego 1462–1476. na wielonurtowy, roztopowy o szerszym korycie. TANG MAOCANG, DONG WENJIE, 1997: Badania wykonano w rejonie Qinghai_Tibetan Infl uences of seven Tibet Plateau raising pro- Plateau, zwanym obszarem źródłowym trzech cesses on climate and environment. Plateau rzek – oprócz wyżej wspomnianych również rze- Meteorology, 16(1): 23–29 (in Chinese). ka Mekong. TODD M., MIKE L., 2003: Natural channel design of step-pool watercourses using the “key-stone” concept. ASCE conference process. 18: 1–11. MS. received April 2010 WANG ZHAOYIN, CUI PENG, WANG RUIYU, 2009a: Mass movements triggered by the Wen- chuan Earthquake and management strategies Authors’ addresses: of quake lakes. International Journal of River Zhaoyin Wang Basin Management. 7(1): 1–12. State Key Laboratory of Hydroscience WANG ZHAOYIN, LEE J.H.W., MELCHING and Engineering, Tsinghua University, C.S., 2009b: Integrated River Training and Beijing 100084, China Management, Springer Verlag and Tsinghua Guo-An Yu Press, Berlin and Beijing. Institute of Geographical Sciences and Natural WANG ZHAO-YIN, MELCHING C.S., DUAN Resources Research XUEHUA DUAN, YU GUO-AN, 2009c: Eco- Chinese Academy of Sciences logical and hydraulic studies of step-pool sys- Beijing 100101, China tems, ASCE Journal of Hydraulic Engineering. Gary Brierley 134(9): 705–717. School of Environment, University of Auckland WANG ZHAOYIN, SHI WENJING, YU Le Liu GUOAN, QI LIJAN, 2009: New challenges of Department of Hydraulic Engineering river management induced by the Wenchuan Tsinghua University Earthquake. Hydroscience and Engineering. 4: Beijing 100084, China 1–11 (in Chinese). e-mail: [email protected]