STUDIA GEOMORPHOLOGICA CARPATHO-BALCANICA

VOL XXXIX ICRAKÓW 2005 PL ISSN 008l iH34

L A N D F O R M E V O L U T I O N I N M O U N T A I N A R E A S

PAWEŁ PROKOP (KRAKÓW)

NATURAL rlAZARDs AND ANTHRopoGENlc IMPACTON ENVIRONMENT IN A TROPICAL MOUNTAIN CATCHMENT,

MEGHAIAYA HILLS, INDIA

Ab§tract. The studies have been carried on in (he Umiew ca(chmeni of 493.7 km2 in the area withhighest rainfal]s in the world reaching 12,000 mm annual]y. Reconstruction of heavy rains, ]andslides,earthquakes was based on various historical sources and collecied rainfall data for the last 150 years.Sate]lite image interpreta(ion and mode]ling soil erosion in a GIS environment supplemented the anal-

ysis. The Umiew catchment encompasses two typical landforms of the southern slope of MeghalayaHills: fores(ed deep canyon and degraded grass covered hilbr pla(eau with small contribu(ion of culti-vable land. Natural hazards combined with human activity have accelerated the environmental degra-dation processes leading, in many places, to complete degradation of vegetation cover and soils. Therelative importance of these processes is controlled by natui.al and anthropogenic factors. The re-

gional continuous heavy rains with 20-25% of average annual precipitaiion in 34 days may causelandsliding on steep but forested canyon slopes. Earthquakes triggered rockfalls and landslides ofwhich material fi]ls river beds up to severa] meters height and cause changes in the river's regime. [nboth cases the absence or presence of forest cover becomes almost negligible. The effects of theseextreme events are long lasting and give rise to high soil losses and the sediment delivery to the rivernetwork. On the con(rary, degraded liilly p]ateau is very resistan( on extreme events. Only locally, agri-cultural land with thicker rego]i[h is susceptib]e to rainfa]l induced shallow landslides. The highest

predic(ed annual soil erosion ra(es up (o 145 t . ha-ł . yr' are [imited to small rields under potato cu]ti-vation on shor( and steep slopes wi(hin hil]y pla(eau. The analysis of maps and satelli[e images over a

period 19]0-1998 shows that the land use/cover in [he study catchment is relatively stable and doesnot a s[raigh[foiward response to the high demographic growth.

Key words: natura] hazards, human impac(, ]ndia

INTRODUCTION

The tropical mountains contain a variety of geosystems sensitive to climaticchanges, natural disasters and socio-economic activity (Al lan 1986; Mes s e rl iand lv e s 1997). Although natural evolution with high frequency of extreme eventscan be an important element of environmental changes (Selby 1974; Starke]

96

1976, 2004), human impact is considered as the dominant driving force of degrada-tion in tropica] areas (Lambi n 1997; Achard et al. 2002). The effects of humanpopiilation pressure leading to deforestation, intensification of agricultural land use,soil erosion, mass wasting, flooding, reservoir sedimentation were observed allover the tropical regions in particular mountains (Ive s and M e s s e r 1 i 1989; Va nLy n d e n and o l d e m a n l997; G e t e and H u rn i 2001 ; V a n a c k e r et al. 2003).

The Meghalaya Hills represents an area where natural hazards (the highestrainfalls in the world, ]ands]ides, earthquakes) combined with human activity(deforestation, surface minżng of limestone and coal, shifting cultivation) have ac-celerated the environmental degradation processes leading in many p]aces tocomplete degradation of vegetation cover and soils (C h a t t e rj e e 1968; S t a r -kel 1972; Ramakrishnan 1992; Starkel and Singh 2004). However scien-[ific exp]oration of Meghalaya Hil]s began in the ha]f of XIX c. (01 d h a m 1854),the knowledge of the complex interrelationships between mentioned factors isfar from complete. Most of the studies concentrated on geology (M a zu m d a r1978,1986) and relief (S t a r ke 1 1972,1989,1996). During last decades ecologistsundertook problems associated with impact of shifting cultivaŁion on environ-ment (To ky and Ra m a k r i s h n a n l983a, b; R a m a k r i s h n a n and K u s h wa -h a 200 l ) and restoration of degraded ecosystems (R a m and R a m a k r i s h n a n1988; Tr i p a t h i et al.1995). On the contrary, only few attempts have been madeto interpret hydrological and soil erosion processes limited to easy accessible ar-eas (Mishra and Ramakrishnan 1983; Froehlich 2004a, b). The applica-tion of remote sensing for estimation of land use/cover changes did not increasedthe recognition of human impact on environment. The rates of deforestation dif-fer considerably between authors due to different methodology (Fo r e s t S u r -vey... 2001; Roy and Tomar 2001; Roy and Joshi 2002). From this review itappears to be a lack in the knowledge concerning interaction between natura]and man-induced processes over long time scales.

The main objective of this study is to evaluate the role of natura] hazards andhuman impact on environmental degradation of tropical catchment over last 150years. More specifically, this study aims at asse§sing the impact of extreme rain-falls and earthquakes on mass movement as well as quantimng their effects to-gether with land use/cover changes on water erosion.

MATERIALS AND METHODS

The annual and monthly rainfall records for the las[ 150 years and daily re-cords for the 15 years 1986-2000 for Cherrapunji and Shillong have been collectedfrom lndia Meteorological Department (IMD) in Pune and Regional Meteorologi-cal Office in Gauhati. Rainfall time series for Sylhet, Mawsynram and Mawphlanghave been collected from different agencies for periods from 50 to 100 years. Onlytwo stations Cherrapunji and Shillong conduct the continuous rainfall measure-

97

men[s. The rainfall intensities have been calculated during 15 minutes periodsfrom hyetographs for both stations for two years 1999 and 2000.

Reconstruction of heavy rains, landslides, earthquakes was based on varioushistorical sources; reports from British administration and field surveys mainly fromXD{ century. Presentday landslides djstribution was estimated during fieldworkand through visual interpretation of False Color Composite (FCC) of Landsat MSS,TM and lndian Remote Sensing (IRS) satellite images for the period 1975-2000.

Land cover data were derived from the Survey of lndia topographic maps1 : 63,360 from 1910, and lndian Remote Sensing (IRS-lD) satellite image withresolution 23 X 23 m from 1998. Maps were transferred into digital form andrectified together with satellite image to the Universal Transverse Mercator co-ordinate system in GIS (ILWIS) environment (International lnstitute...1997) using the 1 : 50,000 topographic maps from 1966 as a target. Training ar-eas for land use/cover map generation from satellite image were chosen dur-ing the 1998-2001 field work using a hand-held GPS. The maximum likelihoodalgorithm was applied for image classification. Four classes of land use/coverwere delimited: forest, grasslands, paddy rice and potato cultivation.

A digital elevation model (DEM) of the study area was created from contoursat intervals of 20 m, digitized from 1 : 50,000 topographic maps. The DEM was in-terpolated to raster with a resolution of 20 m. The proportion of forest in the 200-mclasses of elevation were assessed from DEM and land use/cover data for 1910and 1998.

The Revised Morgan-Morgan-Finney (RMMF) model was used in lIWIS to as-sess soil loss with the help of rainfall data, land use/cover map and calculatedDEM (Morgan 2001). The RMMF model predicts detachment rates by rainfalland runoff. The detachment is compared with the transport capacity of the runoff.The lower of the two values is the annual rate of soil loss.

REsmRCH AREA

The Meghalaya Hills (the name Shillong Plateau is also used) is an active tec-tonic horst, rising up from Miocene time to an elevation of 2,000 m a.s.l., located be-tween the Brahmaputra valley in the north and the Bang]adesh plains in the south(Fig. 1 ). Although geologjcally plateau is a part of Gondwana land, often this area isincluded to the lndian Himalayan Region (IHR), because of the difference to sur-rounding lowlands environmental conditions and human activities (Rao 1994).Hills form the first orographic barrier for the humjd southwest monsoon winds ontheir way from the Bengal Bay. The southern slope at its margin is dissected by steepcanyons up to 1,000 m depth. There are situated Cherrapunji and Mawsynram,which record a mean annuall rainfall of 11,000-12,000 mm. Cherrapunji has theworld record for high rainfalls over durations of between 31 days and two yearssince 1860+}1 (World Meteorological...1994).

98

Fig. ) . Location of Meghalaya Hil]s and Umiew catchmen(. ł -mountajns arid uplands, 2 -isohyeteswith average annua] rainfall [mm] for (he period ]901-2000, 3 -contours every 400 m, 4 -rainfall

stations, 5 - elevations (m a.s.l.)

99

Detailed studies have been carried on in the Umiew catchment of 493.7 km2,encompassing two typical landforms of the southem slope of Meghalaya Hills:deep canyon and hilly plateau. The catchment is drained by the Umiew river80 km long which is the right-bank tributary of the Surma flowing to the Meghna inBangladesh. The river is characterised by strong altitudinal gradients, varying be-tween 60 m a.s.l. and 1,965 m a.s.l. which results in a large variation in meteoro-Iogical, hydrological and ecological conditions over shon distances.

The upper part of the Umiew catchment is built up of Archaean gneisses andquartzites intruded by large granite batholiths (M a z u m d a r 1986). This part hasa mature, structure-controlled landscape with relief energy 50-150 m. The reliefwhich is dissected by the wide valley floors with meandering rivers is the productof long planation and weathering (Bandyopadhay 1972). Metamorphic andigneous rocks are covered near Cherrapunji and Mawsynram by horizontally bed-ded sandstones and limestones of the late Cre[aceous-Palaeogene transgression,several hundreds meters thick. Southern margin of the catchment is truncated bythe Dauki fault and form very steep slope with scarps have 800-1000 m high. Es-carpments are interrupted by a system of deep canyons with amphitheaters ofvalley heads cut into the basement of igneous or metamorphic rocks. The upperpart of their sides is steep, often vertical, and is bordered by complexes of resis-tant sandstones and limestones.

The climate is monsoonal with dły and cool season spanning from Novem-ber to May and warm rainy season from June to October. The mean annual airtemperature (1903-2000) is closely related to elevation and varies between24°C in Sylhet (35 m a.s.I.) and 16.6°C in Shillong (1,500 m a.s.I.). In the most el-evated parts of the plateau, temperatures fall below 0°C in winters thoughsnowfall is rare. The mean annual precipitation (1901-2000) is strongly modi-ried by relief and increases from 4,202 mm in Sylhet to 11,000-12,000 mm in ofCherrapunji (1,313 m a.s.I.) and Mawsynram (1,420 m a.s.l). Than precipitationdecreases with the distance from the edge of plateau to 3,507 mm in Maw-ph]ang (1,840 m a.s.I.) and 2,199 mm in Shillong.

Major soi]s in the area are Ultisols, Alfisols and lnceptisols (Soi] Sur-vey... 1975). The upper part of catchment is covered by old weathered depositsbeing stable for a long time and have given rise to soils up to 2 m deep. Theyoung sedimentary complex around Cherrapunji and Mawsynram has compar-atively less weathered soil material. Soils are generally shallow with dominantdepth ranging from 30 to 50 cm. The soils are mixed with coarser rock frag-ments and the presence of large boulders on the surface indicates the scale ofsoil erosion (P ro k o p 2004). Forest is mostly confined to steep slopes of can-yon and consists of climax subtropical evergreen forest mixed in the lower alti-tudes with tropical semi-evergeen forest and small orange and areca-nut plan-tations (H a r i d a s a n and Ra o 1985). Subtropical pine forest is the secondaryformation growing above 1,400 m a.s.l. It is generally fragmented and scatteredbetween degraded grasslands. Temperate forest (broad-leaved hill forest)

100

which assume covered grassland areas in the past (Bo r 1942) are found onlyin small blocks on hilly plateau above ] ,200 m a.s.l. Agricultural land is limitedto the higher parts of the catchment, whereby the most important crops arepaddy rice and potatoes.

NATURAL HAZARDS

RAINFALLS

The most distinctive feature of the Umiew catchment is the large amount ofrain that annually falls in this area. The knowiedge of rainfall parameters is funda-mental for understanding thresholds of slope stability and the start of overlandflow following soil erosion. For characteristics two stations - Cherrapunji withthe highest and Shillong with the lowest rainfall were chosen.

Trend analysis for annual and seasonal rainfall series shows these to be sta-ble in Cherrapunji and Shillong over at least the last 150 years (Prokop andWa 1 a n u s 2003). But the rainfalls are variable from year to year. The annual rain-falls in Cherrapunji fluctuate between 6,800 mm and 22,760 mm and in Shillongbetween ] ,300 mm and 3,800 mm. The deviations are higher in Cherrapunji com-pared with Shi]Iong and can exceed 1009/o annual average. There is a distinct sea-sonal distribution of rains. The thunderstorms connected with cyclones formedover Bay of Benga] produce large amount of rain between last decade of Marchand May. During monsoon season from June to September 709/o of precipitationfalls. At that time the rain falls almost every day in Cherrapunji and monthly totalsfluctuate between 1,500 mm and 6,670 mm. The highest monthly total duringmonsoon season is only 1,267 mm in Shillong Station has ] 70 rainy days annuallythat is 12 days less compared with Cherrapunji.

The continuous heavy rains and heavy downpours are two types of rains dis-tinguished over northeast lndia (Starkel and Basu 2000; Dhar and Nan-da rg i 2003). They differ in their totals, intensity, duration, distribution in spaceand effects in environmental transformation. This typo]ogy can be also applied torainfall characteristics jn the Umiew catchment.

The continuous heavy rains are caused when the eastem end of the monsoontrough shifts northwards to the Brahmaputra valley or during the periods when"Break" monsoon situations se( in over the lndia with a northward shift of the mon-

soon trough to the foot of the Himalayas (D h a r and N a n d a rg i 2003). Tliese twometeoro]ogical situations are responsible for causing continuous heavy rainfall onabout 65% of occasions (D h a r and N a n d a r g i 2000). The continuous rains last-ing usua]]y 34 days (there have been occasions when rain spe]]s of 6-7 days). Theyhave various intensities and totals exceeding 2,000 mm in Cherrapunji and 500 mmin Shillong, which is 20-25% of annual rain (Fig. 2). These continuous heavy rainscover larger areas from hundreds to thousands square kilometers on the southemslope of Meghalaya Hills.

101

V

TV

®~®

VW V V W

®®

1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

•1 .2 V3 V4Fig. 2. Extreme even(s registered in the Umiew ca(chmenŁ duTing the last 150 years. ] -earth-

quakes, 2 - landslides, 3 - heavy continuous rainfalls in Cherrapunii, 4 - heavy continuousrainfa]ls in Shillong

Orography beside cyclone activity is the additional cause of the enormousheavy downpours in Cherrapunji and Mawsynram (Starke 1 1972; 0 ' Hare1997; S o j a and S i n g h 2004). The highest 24 hours rainfall reached 1,563 mmin Cherrapunji in 1995. The highest rainfall in Shillong was only 415 mm in 1934.But one day rainfall of 381 mm recorded in 1878 was sufficient to cause theflood in Shillong (Sherer 1879). These heavy downpours of duration up toseveral hours are restricted to areas of dozens of square kilometers. They haveespecially high intensity during spring between end of March and May when thecyclone activity is highest.

The knowledge on the intensity of rainfalls is insufficient. In the lndian meteo-rological data publications, the rainfall intensity is calculated at a scale of 24 hours.Hyetograph anabsis for period 1999-2000 shows that the highest rainfa]l during15 minutes reached 53 mm (e.g. 3.5 mm . min-]) in Cherrapunji and 19.5 mm (e.g.1.3 mm . min-') in Shillong. R. S oj a and S. S i n g h (2004) using pluviometers with0.1 second time reso]ution obtained intensities closed to 2.0 mm . min-] during se-veral minutes intervals for the years 1999-2002 in Cherrapunji. Daib rainfalls above500 mm with intensities of 40-70 mm . h-ł appea.r every year in this region.

RAINFALL-INDUCED MASS MOVEMENTS

The Umiew catchment is prone on mass movements due to extreme rain-falls and large contribution (more than 50% of area) of steep slopes above 15°.Landslides are not recorded systematically because a ]arge part of the canyon is

102

no[ accessible during rainy season and its central part is uninhabited. Although itis clear from historical sources that ]andslides were prominent feature.

T. Oldham (1854) mentions the o]dest information about flood (in factlarge hyperconcentrated mud and debris flow) in the Umiew catchment. This oc-curred on the 14 June 1851 and destroyed large portion of Shella village in thecatchment outlet. Rainfall triggered landslides blocked the water flow in therivers. That caused water level rise locally up to 15 meters (it was measured bythe author that the water level rises to a maximum half of this height a duringmonsoon season) and undercut base of the steep slopes. The forested slopeswere scored wi[h gullies and deep ravines, extending from the level of the waterup to the siimmit of the steep slopes. From one of these deep cuts on smal]stream at least 5 thoiisand tones of matter was removed. On the basis ofT. 0ldham's (]854) description it is possible to calculate that torrent wavestarted above Mawphlang and passed along 45 km during 2-3 hours to Shella.That was first noticed catastrophic event since the hil]s have passed under Britishrule in the 1835. Various sources mentioned landslides occurred in 1861,1876,1898 bu[ they never had such effects.

Similar conditions appeared when heavy i-ainfalls and floods occurred innortheast lndia and Bangladesh in 1988. The annual rainfall in Cherrapunjireached 17,925 mm and exceeded the long term average for 63%. The Shillongstation recorded 3,807 mm, the highest annual total for the period 1867-2000. Twoseries of heavy continuous rainfa]ls occurred between 4-7.07. with amounts2,019 mm in Cherrapunji and 522 mm in Shillong as well as between 24-27.08.with amounts 1,989 mm in Cherrapunji and 441 mm in Shillong. The analysis ofrainfalls together with Landsat satelljte images from 1988 allow to find that thewater level rise and connected with it lands]ides are the result of continuousheavy rains of 20-25% of annual average during 34 days. It was clearly visible thatthe Umiew river bed in the deep canyon was filled in sediment carried out fromlandsljdes developed on forested steep slopes. Hence the landslides and not cul-tivated ]and is the main source of sediment delivery to the river network. This typeof extreme events are rare and occur once on tens years (Fig. 2).

The most common mass movements are the shallow landslides and mud-flows. These occur each year but most remain unrecorded. Also they cannot bedelineated on satel]ite images with typica] resolution 20-30 m2 like Landsat or IRS.They usually cover small areas of tens square meters on steep canyon slopes oron the cultivable hilly areas in the northern par[ of the Umiew catchment. Shallowlands]ides and mudflows were observed in [he November 2002 during two daysof continuous rain reaching 270 mm (200 mm fell during 24 hours) in Cherrapunjiwith the highest intensity of 20 mm . h-!. They were restric[ed to very steep slopesabove 35° on natural forested areas. In these cases denuded material very rarelyreaches the rivers. Shallow landslides appeared also near Mawphlang on steepslopes above ]5° on the cultivated land, especially where roads undercut theslopes. Therefore the rainfal] of 250-300 mm during 1-2 days we can treat as the

103

Iower threshold inducing shallow landslides in the Umiew catchment. These con-ditions are met several times during each year, even the years with abnormallylow rainfal] of 7,000 mm in Cherrapunji.

The effects of extreme rainfalls except for rapid water level rise in smallcreeks and boulders sliding on slopes of deforested gullies (S t a r k e 1 1972, 1989)are not visible on grassland hilly plateau near Cherrapunji and Mawsynram. Thedegraded area is wel] protected by a stone pavement. Only removal of the upperlayer of waste cover due to surface limestone, coal and rock mining can give theimpulse to local mass movements.

EARTHQUAKES EFFECTS

Earthquake-induced landslides are an additional complicating factor whentrying to distinguish natural processes from human impact. The seismicity of theMeghalaya Hills is connected with the process of subduction of the lndian platebeneath the Himalaya and tectonics of the Himalayan and Burmese Arcs (K aya 11998). Three severe earthquakes occurred around the Umiew catchment overthe last 150 years in 1869,1897 and 1950 (Fig. 2). The 1869 Cachar earthquake(M = 7.5) is the o]dest noted seismic event (G o dw i n -Au s t i n 1868-1869). It isknown only that it caused the liquefactions and relief changes in the south-east-em part of Meghalaya Hills. The estimated intensitywas 6.0 in Cherrapunji and 6.5in Shillong on the MSK scale (A m b r a s e y s and D o u g 1 a s 2004).

The earthquake of 12 June 1897 in the Shil]ong Plateau is the largest(M = 8.1) intraplate event in the last two centuries occurred in the lndian sub-continent. Much of what is known about the intensity distribution and effects ofthe event comes from R. D. 01 d h a m ' s (1899) detailed report and his subse-quent estimates of three months of heavy rains and aftershock activity follow-ing the earthquake. The earthquake almost totally destroyed settlements andsmall towns on the western part of the Plateau, and caused heavy damage insurrounding areas, chiefly due to the extensive liquefaction, and landslides.Most of these landslides occurred on the southern edge of Shillong Plateau,particularly around Cherrapunji in the Umiew catchment. Geology and relief fa-vour the formation of landslides on steep canyon slopes. The upper part of hillsaround canyon is built up mainly from weathered sandstones and has a muchlower cohesive strength than the underlying crystalline rocks. The verticalsandstone scarps are more easily broken into rockfalls and landslides. On thebasis of R. D. 01 d ha m ' s (1899) old sketch it was possible to estimate that atleast 40yo of vegetation cover from slopes of deep valleys was removed fromcrest to base at many places. The dislodgment of large mass of weatheredrocks and exposure of slopes previously protected by forest caused of supplyenormous volumes of sand to into the river network. The rivers were over-loaded and they channels converted from deep and rocky to shallow filled bysand. The blocking of tributaries caused the water level rise. The highest barri-ers built up with rocks, mud and tress exceeded 60 m high. Accumulated water

104

created great lakes upstream of barriers, which were burst after several days oreven months. That caused the additional landslides on slopes cutting alongriver by flowing water.

Although, deeper into the hills where crystalline rocks dominate and they arenot so deeply weathered as sandstones the landslides were rare and river bedsre[ain their previous character unchanged. Natural effec[s of great earthquake onthe degraded hilly plateau beside enormous building damages in Shillong wereless remarkable and ljmited [o thrown upward into the air stones on hillslopesand smaller landslides induced along the roads.

The impact of landsliding caused by large magnitude earthquakes on sedi-ment budget and denudation over various durations is poorb investigated. Accord-ing to C. F. P a i n and J. M. 8 ow 1 e r (1973) most of the fine grained sediment sup-

plied to river being flushed out within 0.5-2.0 years. A. J. Pearce et al. (1985)found that at least 50-759/o debris retained in the fourth order catchment 50 years af-ter Łhe earthquake in New Zealand. In case of the Umiew catchment large part offine and medium coarse sediment left in the valleys at least during few monsoonseasons. Today it is not visible in the main river, but rounded sandstone boulders of1-2 meter size are still found in the crystal]ine stream beds of four orders.

The secondary effects of great earthquake were the destruction of villages,orange and areca-nu[ groves near the outlet of the Umiew river. The overloaded,shallow river lost capabilities of limestone transport from ]ocal quarries. These

Photo 1. Lower part of the Umiew calchment with developed dense forest coveron ihe devastated slopes 100 years afier earthquake rrom ]897

105

gave the impulse to the migration of people from foothills to higher elevations andlocation of new villages there.

Severa] villages never recovered after the devastation of the earthquake. De-creased population pressure created more "natural" conditions for the regenera-tion of forest cover. The maps from 1910 show vegetation cover developed on for-mer landslides and today they are usually covered by dense forest (Photo 1 ). Thisis the evidence that the relaxation time in the Umiew catchment is similar to otherhumid tropical regions (G a rw o o d et al.1979; Fro e h 1 i c h and S t a r ke 1 1987;Guariguata 1990).

Latest registered severe seismic event (M = 8.5) in the XX century was theearthquake in 1950 with the epicenter on the lndo-China border. Due to remote-ness to the Umiew catchment it was not so remarkable.

The ]4C dates of the palaeoliquefaction features indicate a recurrence periodon the order of 500 yr for large earthquakes in the Shillong Plateau in addition tothe seismic events presented on the Figure 2 (S u k h i j a et al.1999). R. 8 i 1 h a mand P. E n g 1 a n d (2001) connect these events with moderate local earthquakesand large earthquakes in the Bhutan Himalaya. They assume that the Shillong Pla-teau is bounded by two reverse faults in the north and south and the recurrenceinterval for earthquakes resembling th'e 1897 event to be 3-8 kyr on each fault.

ANTHROPOGENIC IMPACT

IAND USĘ/COVER CHANGES

The southern slope of Meghalaya Hills has adequate precipitation, tempera-ture and soil fertility to support vegetation growth (Tr i p a t h i et al. 1995). It is evi-dent that the Umiew catchment was forested before first settling took place i.e.neolithic time (8 o r 1942; 8 a re h 1997). Due to a long cultivation and mining his-tory, with successive periods of land clearing and abandonment, over half of theprimary tropical and subtropical forest has been degraded to grass formations(Ramakrishnan 1992).

Two land use/cover classes -forest and degraded areas covered by grassesoccupied more than 959/o area of the catchment in 1998 (Fig. 3). Time series analy-sis shows very stable proportions between them during last 100 years. Althoughpopulation density grew up from 23 persons . km-2 in 1901 to 123 persons . km-2 in1991 (C e n s u s o f 1 n d i a 199] ), forest area also increased from 44.69/o in 1910 to46.49/o in 1998. Between 1910 and 1998 the rest of larger forest patches disap-

peared on easy accessible areas within hilly plateau. This was partly compen-sated by natural secondary forest regrowth on steeper uninhabited slopes of nar-row canyon. Only few percent of the catchment area converted to agriculturalland use (paddy rice and potatoes) is now cultivated. The paddy rice still coversthe same flat valley bottoms occupied in 1910. It is not possible to compare in de-tail changes of potato cultivation on steep slopes. The fields are usually very small

106

Fig. 3. Land use/cover map of the Umiew catchment compiled on the basis of satellite image]RS-lD from 1998.1 -forest, 2 -paddy rice in larger valleys, 3 -potato cultivation on slo-

pes, 4 - degraded grasslands/rock outcrops

and intermix with fallow land or degraded grasslands. Due to that they have notbeen delineated on the maps.

Forest distribution varies with e]evation (Fig. 4). Generalb, the proporiion offorest fluctuates between 80-95% up to 1,200 m a.s.l. Above this a]titude it con-stantly decreases to few percent at highest elevations. This distribution is closev re-Iated to relief and population density. The lower part of the catchment, below 400 ma.s.I„ has been partly deforested due to intensive surface limestone mining frompre-colonial times or location orange and areca-nut plantations. The central part of

107

m a.s.l.

2,000

1 ,800

1,600

1 ,400

1 ,200

1 ,000

800

600

400

200

00 20 40 60 80 100

[%]Fig. 4. Changes of the proportion of forest area wilh the elevation in 19]0 and ]998

in the Umiew catchment

the canyon is severev dissected a dense drainage system about the density ofabove 10 lm . km-2. The forest proportion exceeds here 959/o on steep slopes butdisturbance has been increasing over the last decades as a selective logging andfuelwood extraction (Roy and Tomar 2001). Contribution of the flat areas in-creases above the 1,200 m a.s.l. They are usually severev degraded in the e]eva-tions of 1,200-1,600 m a.s.I. The upper part of the Umiew catchment has the highestpopulation density, above 350 persons . km-2 and the lowest forest proportion ofabout 79/o. The settled agriculture developed here to meet commercial needs ofShillong township. This is also area where are the largest changes of forest spatialdistribution are found. The broad-leaved forests were converted to pure pine for-ests. The second one is converted to agricultural land use connected with s]ash andburn agriculture (Mishra and Ramakrishnan 1983) or it is utilised for com-mercial purposes. These processes result in forest ffagmentation. Local authoritiesmainb due to land tenure system do not undertake the afforestation action. Most ofthe forests are in the private hands or village communities. Although fragmentationproceeds on flat areas the small patches of forest are still kept by farmers. In manycases young trees are left on fallow land for a growing. This is essential e]ement ofslash and burn agricultural system in higher elevations of the Umiew catchment(Tiwari 2003).

108

SOIL EROSION

lt is generally assumed that land use/cover changes have caused acceler-ated soil erosion and increased sediment yields in the rivers (M o r g a n 1986). Ex-treme rainfall and steep slopes in the Umiew catchment create potential condi-tions for efficient soil erosion. There are only two micro-scale erosion measure-ments close to the investigated area. Both experiments gjve an important infor-mation about the scale of erosion from two contrasting land use types on hillslopeplateu: potato cultivation and degraded grasslands.

Soil loss was found to range between 34 and 56 t . ha-' . yr-` for potato cultiva-tion at 40° steep slopes near shillong (M i s h r a and R a m a k r i s h n a n l 983). Sig-nificantly lower values were ob[ained for fallow land -only 7 t . ha-t . yrł. It wasalso found that with increasing of fallow duration runoff and soil ]oss were furthersignificantly reduced to 2 t . ha-! . yr].

]37Cs and 2]°Pbex techniques were used for the study of soil loss in the grassland

hillslopes near Chenapunji (F r o e h 1 i c h 2004a, b). This study has shown that pres-ent rates of soil erosion are onv 2.1 t . ha-[ . yr-' and most of the sediment from de-graded grasslands is deposited at the foo[ of the slopes. Stream and recent floodplain sedimen[s are derived primariN from gulv and channel bank erosion.

Described environmental conditions near Cherrapunji are representative fora half of the Umiew catchment (Photo 2). The deforestation in the past caused ex-cessive overland flow that removed fine sediment and produce armoring debris

Photo 2. Deforested hilly plateau wi[h degraded grasslands near Cherrapunji

109

with hard impermeable layer on soil surface which range up to 10-20 cm. Sandsand gravels constitute more than 60% of the upper soil profile (Pro kop 2004).Thesoilsareheavilydegradedandquiteoftendevoidiipperhorizonsandinmanyplaces they have a character of a waste cover only.

The application of the RMMF model complete soil erosion pattem in the en-tire catchment. Annual soil erosion predicted by model is moderate and rangefrom 0-145 t . ha-` . yr-`. Two contrastive land use/cover types, natural dense forestand degraded due to human activity grasslands, give similar response to the soilerosion rates. The soil loss is the lowest in the valley bottoms under rice cultiva-tion, dense forest and degraded grasslands -below 10 t . ha-' . yr' (Fig. 3). Thoseare within the tolerance limits of the mountain landscape (M org a n 1986). Thehighest erosion rates up to 145 t . ha-] . yr] (average 60 t . ha-] . yrł) are under po-tato cultivation on short and steep slopes within hilly plateau in the northern partof the Umiew catchment.

The FCC satellite image analysis for the period 1975-2000 show that theUmiew river bed in the deep canyon is not filled in sediment during years with av-erage rainfall. This is the evidence that most of the mobilized sediment from culti-vated fields is being trapped on the flat and wide valley bottoms in the area withmature relief. Their favour to this terracing rice fields. Similarly, only little hillslopesediment is delivered to the streams in the area with degraded grasslands. Thesediment delivery to the Umiew river bed dramatically increases during extremeevents like heavy continuous rainfalls and severe earthquakes induce large lands-lides that give rise to high soi] Iosses.

CONCLUSIONS

The area of the Umiew catchment is characterised by intense geomorphicprocesses including mass wasting, water erosion and fluvial activity. The relativeimportance of these processes is controlled by natural and anthropogenic factors.Natural extreme events are the dominant driving forces in environmental degra-dation of the Umiew catchment. The role of the human impact depends on the ]o-cal environmental conditions and land use history.

The area of the canyon with steep slopes is most sensitive on mass move-ments (rockfalls, landslides and debris flows) triggered by extreme rainfalls andearthquakes. The regional continuous heavy rains with 20-25yo of annual precipi-tation in 3-4 days may give to water level rise and cause landsliding on steep butforested slopes. Severe earthquakes may trigger rockfalls and landslides of whichmaterial fills river beds up to few meters high and changes river's channel regime.The absence or presence of forest cover in both cases becomes almost negligi-ble. The majority of the sediment delivered to the river system is produced bymass wasting on forested land during these extreme events. The effects are longlasting and give rise to high soil losses for a number of years.

110

The hilly p]ateau is under strong influence of human activity. The relative roleof human impact depends on the regolith thickness, the relief, and land use historyof the area. The most degraded grasslands are very resistant on extreme events dueto stone pavement on the soil surface. The highest soil erosion rates were in thepast and presentday hillslope erośion is very low. The ]andslides are not obseived.Only heavy storms can generate sediment through gu]ł and stream bank erosion.On the contrary, cultivable land with thicker weathered cover is prone on rainfall in-duced landslides and that annual soil erosion rates are highest here.

Both contrasting landfoms and connected with them land use/cover types -canyon with natual dense forest and degraded by human activity grasslands withinhilb plateau give similar low soil erosion rates due to tota]b different reasons.

The analysis of maps and satellite image over a period (1910-1998) show thatland usdcover in the study catchment is relativeb stab]e and does not a straightfor-ward response to the high demographic growth. The major ]and cover change pro-cesses observed are expansion of grasslands-croplands and reforestation with co-niferous forest, leading to a more fragmented landscape structure in the upper partof the Umiew catchment. The deforestation in the uplands is compensated for bya regeneration of secondary fores{ on abandoned low-lring areas.

ACKNOWLEDGEMENTS

This paper was completed as part of an lndian National Science Academy(INSA) fellowship and cooperation with the Department of Geography North-East-ern Hill University (NEHU) in Shillong under the scientific exchange programmebetween lndian Department of Science and Technology (DST) and Polish Com-mittee for Scientific Research (KBN).

Departmen[ of C;eomorphology and Ilydrology of Moun[ains and Uplaridslnstitute of C;eography and Spatial Organiz,ation, Polish Academy of Sciencesul. Śii). Jana 223IJ)IS K].ahóu), Polande-mail: pau)el@z8.pan.krakow.pl

REFERENCES

Achard F., Eva H. D„ Stibig H. J., Mayaux P, Gallego J., Richards T., Malingreau J. P.,2002. Determination of Defores[aiion Rates of ihe World's Humid Tropical Forests . Sdience 297 ,999-1002.

Al la r\ N . J .R ., \986. Accessibiliy and allitudina[ zonation model of mountains. Mour\\ti\n ReseaLrchand Development 6, 185-194.

Arr\braLseysN.N.,Doug1as].,2004.MagnitudecalibrationofnortiTlndianearthquahes.Geophys-ical Journal fn!ernational 159, 165-206.

8 aindy opadhay M. K.,19]2. Geomorphological Charac[erislics of lhe Southern Part oi the KhasiJJi.//s. Geographical Review of lndia 34, 2, 184-189.

BaLreh H., \997. The his[ory and culture of lhe Khasi people. De"\, 5\2 pp.

m"

8 i 1 h a m R., E n g 1 a n d P„ 200 ] . P/a/ecw `pop-up' i.n tłm grć'crf /897 Assam earfhquaA€. Nature 4 ] 0,806*09.

8 o r N. L., ] 942. RĆ'//.c/ ŁJeg€/a/i.on o/Sh/.//ong fya/eaŁ/-Assam. Indian Forest Records 3, 6, 152-] 95.Census of lndia,1991. New Delhi.Ch a\\ e r'ie e S. P„ 1968. La geographie resionale du Pla[eau de Meshcilaya, |.in..l lndia,. Resional S[u-

dł.cs, ed. R. L. Singh, IGU, Calcutta, 218-244.D h a\r 0. N., N ain d aT g.i S., 2000. A Study oi Floods in the Brahmaputra Basin in lndia. lnlernatiional

Journal of Climatology 20, 77]-781.D h aL r 0. N., NaT\d aT g.i S., 2003. Hydrometeorological Aspects of Floods in lndia . Na\uTa\ HazaLids 28,

1-33.

Forest Survey of lnclia, 2001. rtłe S/a/Ć. o/rones/ J?ć'porf. Dehra Dun, ]30 pp.F: r o ehl'i c h W„ 2004a. Soil erosion in experimenial ca[chment , Li\r\..| Rainfall, runoff and soil erosion

in the globally ex[reme humid area, Cherrapunji region, India, eds L. S`2irkel, S. stingh, PraceGeograriczne ]9], 8]Ę9.

Fr oehl.ich W., 2004b. Soil erosion, suspended sed.iment sources and deposi[ion in the MaLu-Ki-Sy-iem drainage basin, Cherrapunji, noriheastern lndia, |.rT\..| Sedimen[ Transfer Through The Flu-Ł)ł.a/ Sys/gm, eds V Golosov, V Belyaev, D. E. Walling, IAHS Publ. 288,138-]46.

F r o e h 1.i c h \N., S t Ęir ke 1 L„ \987 . Normal and ex[reme monsoon rains -[heir role in the shaping ofłł]Ć. Da7.Ć'ć'/ł.ng H/.ma/c7ya. Studia Geomorphologica Carpatho-Balcanica 21,129-160.

G ar w o o d N. C., J an o s D. P., BT okaLw N.,1979. Earthquake-Caused Lands[ides.. A Maior Dislurb-ancć' /o rn?pł.co/ Fonć's/s. Science 205, 997-999.

G e \e Z., Hur T\i H„ 2001. Implica[ions of land iise and land coi)er dynamics for mountain resourcedegradaiion in [he Nor[hwes[em E[hiophlan highlands. Mour\\tiin Reseaich and Developmen\21,184-191.

Godw.in-A,us\er\, \&68-\8&9. Ear[hquake in [he Cachar Hills, Exiracts h.om Le[ters from Cap[ainGodŁŁ;/.n-AŁłs/Ć'n. Proceedings of the Royal Geographical Sociev of London ]3, 5, 370-372.

Ciuar.igua\a M. R.` 1990. Landslide disturbance and fores[ regeneralion in the Upper Luquillo Mo-Ł/r}/cJł.ns o/ ftłer/o Rł.co. Journal of Ecology 78, 3, 814Ę32.

H a r i d a s a n K., R a o P. R., ] 985. for€sf #o/a o/Mcgha/aya. Vol.1,11, Bishen Singh, Dehra Dun, 905 pp.International Jnstitute for Aerospace Survey and Earlh Scienc.es,1997./ŁWZ5-

The Jn[egra[ed Land Wa[er lniormaiion Syslem. User's Guide. Er\schede, The Neiherlands,511 pp.

Iv e s J. D`, M e s s e r l.i 8.,1989. The Himalayan Dilemma: Reconciling Deuelopmen[ and Conseruaii-on. Routledge and UNU Press, London, New York, 296 pp.

Kay al ]. F`.,1998. Seismicib/ of norlheast lndia and surroundinss -development oiJer [he past 100yea#. Joumal of Geophysics 19, 9-34.

L.aTnb.\n 8. F.` 1997 . Modelling and moni[oririg land couer charige processes in lropical resioris. Pro-

gress in Physical Geography 21, 375-393.M aLzumdaT S. K., \978. Morphogene[ic Euolu[ion ol lhe Khasi Hills, Megha[aya. MisceH)aineous P`i-

blications 30, Geological Survey of lndia, 208L213.M aLzu rri d ar S. K„ \9&6. The Precambrian h.ameu)ork of part of lhe Khasi Hills, Meghalaya. F+ecoids

of lhe Geological Survey of lndia 1 ]7, 2,1-59.M e s s e r 1 i 8., lv e s J. D. (eds),1997. MocJn/cr/.ns o/fł}Ć> Wor/d.. A G/oba/f}r.o/J.o;. Pai'thenon Publishing,

New York, London, 493 pp.M i s h ra 8. K., R a m a k r i s h n a n P. S„ 1983. S/czsh and bŁłm qgn.cŁ//rŁw€ a/ higher ć'/€t)cm.ons /.n Wor-

{h-Eastem lndia. 1. Sedimen[, u)ater and nutrien[ Iosses. Agi-ic"uTe, Eicosystems and Enw.ironment9, 69Ę2.

Morgan R. P. C„ 1986. So/./ erosł.on and consć'ma/ł.on. Longman, 298 pp.Mor gan R. P. C., 2001. A simple approach lo soil loss prediction: a reuised Morgan-Morgan-Finney

mode/. Catena 44, 305-322.0` Hare G., ]997. 71hc /ndł.an monsoon, Part 2: 7lhe rał.ns. Geography 82, 335-352.

112

01 d h a m R. D., 1899. Reporf on iłm gmaf Ć'anrigHa*ć. o/ /2`" Jcłn€ /®7. Memoirs of the GeologicalSurvey of ]ndia 30, Calcutta, 379 pp.

01 dh aLm T.,1854. On the geologica! s[riicture of part of [he Khasi Hills. Ber\8aL\ Memoirs o[ \he Geo-logical Survey of lndia 1, 2, 76 pp.

Pa.in C. F., Bowler ]. M., \973. Denuda[ion following lhe Nouember ]970 ear[hquake al Madang,PiqpŁła NCŁŁJ CŁ/riea. Zeitschrift fur Geomorphology, Supplemen[ Band 18, 91-104.

P e ar c e A. ] ., 0 ' 1. o u ghl.\ r\ C. l..,Wal s o n A. J.,1985. Medium-term effiects oi landsliding and relat-ed sedimentation eualuated liiily years afler an M7.7 earthquake. lTnerT+aŃioT\al SympostiHm.. De-bris #oŁt)s and dł.sas/€r f}€#n/i.on, September 315, 1985, Tsukuba, Japan, 291-296.

P r o ko p P., 2004. The Soils, |.m..| Rainfall, runofl and soil erosion in the globally extreme humid area,C7?cm7pzJ"./. rć'gi.on, /ndi.a, eds L. Starkel, S. Singh, Prace Geograficzne ]9] , 4243.

P r okop P .,\^1 aliir\us ^, T008. Trend and periodicity in the longest irislnimental rainf:dll series for theareaofmostexmemerainfalliniheu}orld,northeas[India.Geogiaphiaporiori+cal6,2,25*;5.

R am S. C., R`aLm aikr .\shn aLn P. S.,1988. Hydrolosy and soil feriility of degraded grasslands at Cher-rapŁłn/'J. ł.n JVo"h-Ea5/gm /ndł.a. Environmental Conservation 15, 28L35.

Ra m akr.i sh n a n P. S.,1992. Shif[ing Agriculture and Sustainable Deuelompent: An lnierdisciplinaryS/Ł/dy ńiDm Nortłi-Easłem /ndi.ci. UNESCO-MAB Series, Parthenon Publications, Paris, 422 pp.

R a m a k r i s h n a n P. S., K u s hw a h a S. P. S., 2001. Seconć7ciw /ores/s o/ fhć' łJi.ma/aya ŁŁ)ł.fł} cm-

phasis on the north-eastern hill region of lndia. ]ournaL\ o[ Trop.ica\ Fores\ Sciience 13, 4,7 TJ -7 4J .

F\aLo K. S.,1997 . Natural Resource Managemen[ and Development in Himalaya. A Recourse [o lssuesand S/ria/Ć'gł.cs. ENVIS Monograph 1, GB Pant lnstitute of Himalayan Environment and Deve-lopment, Almora, ]ndia, 56 pp.

Roy P. S., ]ostri P. K., 2002. Forest couer assessment in nor[h-eas[ Jndia -the potentia[ of [emporalu)ide su)alh salellite sensor data (lRS-] C WiFS). lnlerT\ati\or\a\ Journa\ o[ Rerr\o`e Senstng 23, 2,48814896.

R o y P. S., To m a r S., 200 ] . Łandscape cotJer dyncmł.cs paffć'm ł.n A4€gha/aycr. Interna[ional joumal ofRemote Sensing 22, ]8, 3813-3825.

S e 1 by M. J., 1974. Domł.nanf geomophi.c et)en/s ri /and/orm cc)o/Łł/r.on. Bulletin lnternational Assoc-iation Engineering Geology 9, 85i9.

Sher er ]. F., l&R9. Repor[ on [he adminis[ra[ion of [he Khasi and Jain[la Hills for lhe year 1878-79.CalcuŁŁa, 14 pp.

Soi I Su rvey S taff,1998. Ke)Js /o Soi./ TCŁronomy. United States Department of Agriculture, Was-hington D.C., 332 pp.

S o.] a R., S.ir\gh S., 2004. The rainfall characteristics, |.in..| Rainiall, runoff and soil erosion in the glo-bally extreme humid area, Cherrapunji region, India, eds L. S\eirke\, S. S.iT\gh, P{ace Geogra.fii-czne 191, 59-71.

S\arkel 1..,1972. The rnodelling of morNsoon areas of lndia as related to catas[rophic rainfalls. Geo-

graphia Polonica 23, 151-173.S t a`rkel L., \976. The role of extreme (catastrophic) meteorological euerils in contemporary euolut~

ł.on o/s/opes, [in: ] GĆ'omopho/ogy andc/ł.mcr/g, ed. E. Derbyshire, J.Wiley, Chichester, 203-246.S\arkel 1.., \9819. Valley floor euolu[ion in [he marginal areas of the Himalaya mouniains and the

Kłłcłsł.-Jał.nfi.a f?/crłgcru. Zei[schrif! fur Geomorphology, Supplement Band 76, IĘ.S t aLT ke 1 L„ \996. Ptesenl-day Eormation of the Southem Part of Meshalaya Plateau. H.ill Geographer

]2' ]-2, ]3-]9.S i arkel L., Ż004. Temporal Clustering oi Extreme Rainfall Euen[s in Reliei Transformation . Jourr\al

Geological Society of lndia 64, 4, 517-523.StaLrke\ L., 8 aLsu S. Ceds.), 2000. Rabis, Lcindslides and Floods in lhe Datieeling Himalaya. lntiaLn

National Science Academy, New Delhi, 168 pp.S`a`rkel L`, S.ingh S. (eds.), Ż004. Rainiall, runoif and soil erosion in the g[obally extrerne humid

anea, Cherrapunji region, India. Pra`ce Geog{aLfi\czne 191,110 pp.

113

Sukhija 8. S., Rao M. N., Reddy D. V, Nagabhushanam P, Hussain S., Chadha R. K.,G u pta H. K., \999. Timins and relurn period of major paleoseismic ei)en[s in the Shilloru] Pla-/eau, /ndł.a. Tectonophysics 308, 53i5.

T.iw air.i 8. K., 2003. Innoi)a[ions in Shiiting Cultii)ation and Land Use and Couer Change in HigherEleuations oi Meghalaya, India, [.m..| Methodological issues in mouniain research. A socio-ecol-ogł.ca/ sys/em appmcrch obsema/ł.ons, eds P S. Ramakrishnan, K. G. Saxena, S. Patnaik, S. Singh,UNESCO, MAB, lcIMOD, NEHU, New Delhi, Oxford and IBH Publishing Co. Pvt. Ltd.,163-175.

To k y 0. P, R a m a k r i s h n a n P. S., 1983a. Secondary sŁłccessł.on /o//oŁoł.ng s/ash and bŁm agn.cŁ///u-rehNorth-Eastemlndia.I.Biomass,Iiiterfallandproduc[iuity.Jciurnalo[Eico+ogy71,735=]45.

To k y 0. P, R a m a k r i s h n a n P S., 1983b. SĆ'condaT sŁ/cccss/.on /o//oŁŁJJ.ng s/ash and bum agrJ.cŁł/m-i€ in North-Eas[em lndia. 2. Nutrieni Cycling. ]oumail o[ Ecohogy 71, 747-]57

Tr i p a t h i R. S, S h a n k a r U., Pa n d e y H. N., 1995. f}iesen/S/ałus andsłmłggł.es /orEcorć's/orc7fł.on o/Degraded Ecosystem a[ Cherrapuriji, |im:.| Ecores[oration of Degraded I]ills , eds 8. K. TiTwaLrii,S. Singh, Kaushal Publ., Shillong, 23-59.

Va n a c k e r V., G o v e r s G., 8 a r r o s S., P o e s e n J., D e c k e r s J., 2003. rhe e//gc/s o/shorf-ferm so-cio€conomic and demographic change on landuse dynamics and i[s corresponding geomorp-hic response u)ilh relation [o u)aler erosion in a tropical mountainous ca[chment, Equador.Landscape Ecology 18, 1-15.

VaL r\ Ly T\ d e n G. W. ]., 01 d e m ain R.1.., \997. The Assessment oi [he Sta[us of Human~Induced SoilDĆ'grodc7/i.c)n ri SoŁłfł7 Ć7nd so«4heas/ As/.a. ]nternational Soil Rererence and lnformation Centre,Wageningen, 41 pp.

Wo r 1 d M e t e o r o 1 o g i c a I 0 r g a n i s a [ i o n, 1994. Cuł.de /o fJydno/ogł.ca/ f}crc/ł.ces. 5th edn., 168, Ge-neva, Switzerland, 735 pp.

STRESZCZENIE

P. Prokop

NATURALNE ZAGROżENIA I WPŁYW CZŁOWIEKA NA ŚRODOWISKO W GÓRSKIEJ ZLEWNI

TROPIKALNEJ, WYŻYNA MEGHAIAYA, lNDIE

Badania prowadzono w latach 1999-2002 w zlewni Umiew o powierzchni 493,7 km2, obszarzeo najwyższych na świecie opadach, sięgających średnio 12 000 mm rocznie. Zlewnia obejmuje dwacharak[erystyczne dla południowego skłonu Wyżyny Meghalaya typy rzeźby: kanion porośnięty tro-

pikalnym i subtropikalnym lasem oraz pagórkowa[e plateau porośnięte zdegradowanymi formacjamitraw z niewielkim udziałem terenów użytkowanych rolniczo. Naturalne zdarzenia ekstrema]ne są do-minującą siłą prowadzącą do zmian w Środowisku badanej zlewni. Wpływ człowieka ma znaczeniedrugorzędne i zależy od warunków lokalnych oraz historycznych uwarunkowań. Stwierdzono, że roz-lewne opady sięgające 20-25% średniej sumy rocznej oraz trzęsienia ziemi powyżej 7 stopni w skaliRichlera przyczyniają się w największym stopniu do powstania ruchów masowych i degradacji szatyroś]innej. Osuwiska rozwinięte na za]esionych stokach są głównym źi.ódłem dostawy materiału dosieci rzecznej. Pagórkowate plateau, gdzje faza największej erozji wystąpiła w przeszłości, jes[ bardzoodporne na zdarzenia ekstremalne. Związane jesi to z występowaniem na powierzchni silnie sce-mentowanego, szkieletowego bruku ograniczaiącego procesy grawitacyjne i erozję. Jedynie północ-na część zlewni z grubszą warsh^rą zwietrzeliny jest narażona na pvtkie osuwiska oraz występującąlokalnie wysoką erozję do 145 t . ha-' . rok-' z upraw okopowych. Ponad pięciokrotny wzros[ gęstościzaludnienia w zlewni Umiew w XX wieku doprowadził wprawdzie do większej fragmentacji lasówna pagórkowat)m plateau, równocześnie jednal( wzrosła powierzchnia leśna w obrębie opusz-czonych przez ludność stromych s[oków kanionu.