Research on the processes of erosion and sedimentation, and their effects from human activities in onga river basin, kyushu, japan

Thus, the thesis focuses on long-term natural processes such as sedimentation and erosion through a case study in Onga River Basin in northern Kyushu, Japan to clarify a sediment storage

RESEARCH ON THE PROCESSES OF EROSION AND SEDIMENTATION, AND THEIR EFFECTS FROM HUMAN ACTIVITIES IN ONGA RIVER BASIN,

KYUSHU, JAPAN

Tran Anh Tu

HUMAN ACTIVITIES IN ONGA RIVER BASIN,

KYUSHU, JAPAN

A Thesis Submitted

In Partial Fulfillment of the Requirements

For the Degree of

Doctor of Engineering

By

Tran Anh Tu

to the DEPARTMENT OF CIVIL AND STRUCTURAL ENGINEERING

GRADUATE SCHOOL OF ENGINEERING

KYUSHU UNIVERSITY

Fukuoka, Japan August, 2011

Fukuoka, Japan

CERTIFICATE

The undersigned hereby certify that they have read and

recommended to the Graduate School of Engineering for the acceptance

of this thesis entitled, ‘‘Research on the Processes of Erosion and

Sedimentation, and Their Effects from Human Activities in Onga River Basin, Kyushu, Japan’’ by Tran Anh Tu in partial fulfillment of

the requirements for the degree of Doctor of Engineering

The amount of sediment discharge in Japanese rivers is showing a very high value in the world This is considered to be affected not only by natural processes such as landside and flooding under humid subtropical climates after the last glacial period but also by human activities such as agriculture and mining For management and prediction of watershed sediment, it is important to know its long-term sedimentation and relative factors Thus, the thesis focuses on long-term natural processes such as sedimentation and erosion through a case study in Onga River Basin in northern Kyushu, Japan to clarify a sediment storage mechanism in the Japanese catchment where many human activities since late Holocene were detected.The contents of this dissertation are presented as follows

Chapter 1 introduces the overview of research background and objectives It is focused on the sedimentation and erosion for some thousands of years for the drainage basin For a long-term period, part of sediment lost due to discharge out of the basin or to the sea It is needed to concern on all processes in systematically such

as erosion, sedimentation and sediment delivery ration This chapter also mentions to overview of the comparison between sediment yield without and with human activities

Chapter 2 mentions the study area characteristics and methodology to calculate the sediment erosion and deposition for long-term and short-term Characteristics of the basin focus on the geological units which were formed in Quaternary periods, especially in Holocene The sea level changes in Late Quaternary are considered in relation to the marine sediment formed in this time such as Onga silt layer which contained shell mounds and tephra These records are the markers to identify the relative age of the silt layer for analysis in chapter 3 The method is used to estimate long-term erosion or denudation which is the function of mean altitude is described Long-term erosion calculation was also adopted in Japanese conditions as the function of altitude of dispersion based on the data of sediments stored in the reservoirs throughout Japan For short-term erosion estimation, among some models, RUSLE is proposed to use The parameters of this model were also modified and applied in some type of land-cover and landscape such as for farmlands and forests

in steep hill slope by many researches

Chapter 3 explains in detail the processes to construct the palaeo-surfaces At

bedrock is the boundary For Holocene and Pleistocene boundary, based on the characteristics of geological units and sea level changes as mentioned in chapter 2, the boundary is defined when upper most silt layer change to coarser grain size the changes to the bottom To support this viewpoint, some cross sections, shell mounds, volcanic ash, geotechnical parameters and sediment comparisons are also used to analyze in accompany with literatures Beside boreholes, geological map and DEM are used Differences in resolution between DEM and geological map cause the errors in some parts of the study area such as the Holocene geological boundaries pass through the hill or mountain on DEM The existed geological map in scale 1:200.000 needs to modify to adopt with DEM 50m resolution The altitudes of lower Holocene formations are interpolated to make continuous palaeo-surface Among some interpolated methods supported in GIS, Spline with tension is the suitable method to do this work It can give a ridge or valley, and the output is exacted with the input in location of boreholes Errors from interpolation are defined

by checking the thickness of sediment in comparison with maximum thickness in the nearby boreholes Errors are also modified based on the U-shape of the ancient channels in lower reaches of the river

Chapter 4 concentrates in the analysis of sediment storage and erosion in long-term such as in Holocene and Quaternary The volume storages are summed up into the catchment 5, 6 7 and 8 and compares to its catchment area The results show

a proportional correlation between sediment volume and its area The annual average deposition in Holocene was higher 30 times than that was in Late Pleistocene From the sediment storage, the average sedimentation rates in the basin are about 1.6 mm/year, and maximum is 2.0 mm/year, and average erosion rate varies from 0.114

to 0.163 mm/year From the sediment budget result, the SDR in Holocene (8500 year BP) of the Onga River varies from 0.0 to 0.1 A-values are higher in Tagawa than orthers, while deposition in Iizuka region is higher In Tagawa region the sediment discharge ratio is higher than other regions For examples, for catchments order 6, minimum SDRs vary from 0.59 to 0.92, especially, these values can be up to 0.82 (Tagawa 1) and 0.92 (Tagawa 3) SDR high can cause some issues relate to the water construction below the catchment as fast filling the dams

calculate the ability of soil loss (A) in 1997 The product of R and LS is the RULSE’s natural parameters which can be compared with the erosion from Ohmori’s model RLS consists of 48 % of total erosion from Ohmori’s model It means the sheet and rill erosion consists of 48% of total erosion in the catchment On the other hand, 52% of total soil loss is from other sources such as slope failures which may deposit in hill slope I Holocene, in system of total erosion, sediment storage in channels, sediment storage in hill slope, surface erosion and sediment discharge out

of the Onga River basin, at least 48% of sediment is yielded from surface erosion can

be reach to the channel, and then (SDR) 10% of sediment in the channel can be discharge out of the basin Beside natural processes, human impacts also are classified by catchments The impacts are clear in the catchment 6 and 7 In catchment with stream order 5, there are also recognized the sediment caused by human activities but in some area with slope angle over 13 degrees, the natural sediment yield is dominant A-values are near the E values because most of these areas locate in near/in the forest Sediment erosion (A) from agriculture and forestry

is about 69000±10% ton of soil loss per year higher than natural-base level

Chapter 6 summarizes the content of the research

With utmost sincerity and pleasure, I express my profound gratitude to my mentor Associate Prof Yasuhiro Mitani for his supervision, guidance, critical comments He gives a view of comprehensive and systematic research on the study issues His understanding and full supporting outside the study were very important for daily life in Japan

I would like to thank Prof Noriyuki Yasufuku and Prof Yukihiro Shimatani who are the members of the Doctoral Examining Committee for their valuable comments and suggestions to complete this dissertation

I express my sincere thanks to Prof Koichiro Watanabe for his recommendation

tothis Course I would like to thank Japan International Cooperation Agency (JICA) and ASEAN University Network/ Southeast Asia Engineering Education Development Network (AUN/SEED-Net) for supporting me the scholarship to complete Special International Doctor Course During the time I study and research

in Japan, JICA have not only thoughtfully cared me and my family in daily life but also encouraged me in research

My deeply thank go to Assistant Prof Hiro Ikemi for the patience and continuous support through all the research work It is also thanks to other members

in the Environmental Geo-technology lab, who help me enthusiastically for daily communications in Japanese

I would also like to thank the Board of Rector of Ho Chi Minh University of Technology who gave me a permission to attend this Course, and collogues who have supported and shared my works during the time my research has been doing in Japan

I cannot find words to express my gratitude to my parents for their love and support To my wife, Thuy Duong, I extend my loved warmest and deepest thanks for her great patience and endless support Through the early morning, long evenings, and overnight lab working, she has stood by me and gives helpful encouragement

My lovely two children, Tu Han and Nhat Nam, are the greatest encouragement

2.2 Method to estimate sedimentation and erosion 13

2.2.1 Method to estimate sedimentation 13 2.2.2 Method to estimate erosion 14

2.3.1 Universal Soil Loss Equation 16

2.3.2 Water Erosion Prediction Project Model 23

3.2.3 Interpolation methods 43 3.2.4 Processes to reconstruct the palaeo-surfaces 48

4.1.1 Sediment storage in Quaternary 56

4.1.2 Sediment storage in Holocene 58

4.2.1 Erosion rate from sediment storage 63

4.2.2 Estimating erosion from long-term erosion model 63

5.1.1 Rainfall and runoff erosivity factor 73

5.1.2 Soil erodibility factor 76

5.1.4 Slope steepness factor 82 5.1.5 Cover and practice factors 84 5.2 Erosion in forest areas 86

5.3 Soil loss from agriculture and forestry activities 89

References

Figure 2-2 Geological map of Onga River basin 10 Figure 2-3 Cross section along Onga River 11

Figure 2-4 Relative sea level change in the Japan from stage 5e to present 12

Figure 2-5 Example calculation of thickness from two DEMs in GIS 14

Figure 2-6 Schematic of a small watershed which the WEPP erosion model 23

Figure 2-7 Units conversion graph 25 Figure 3-1 DEM 50m displaces the Onga River basin 33

Figure 3-2 Quaternary geological units in Onga River basin 34

Figure 3-3 Distribution of boreholes 35 Figure 3-4 Flow chart to extract the geological boundary 36

Figure 3-5 Statistic of number of points change elevation from hill slope to gentle

Figure 3-6 Q-T boundary is extracted from DEM 38

Figure 3-7 Modified Holocene boundary 38 Figure 3-8 Location of cross section A-A’ 40

Figure 3-9 Cross section of Quaternary sediment in Onga town (A-A’) 40

Figure 3-10 Typical sediment grain sizes in the study area 41

Figure 3-11 Attributes of boreholes within Holocene boundary 42

Figure 3-12 Flow chart is used to define the Stream network, watershed and stream

order 43

Figure 3-13 Surface is interpolated by Spline with tension method 46

Figure 3-14 Surface is interpolated by IDW method 47

Figure 3-15 Surface is interpolated by Kriging method 47

Figure 3-16 Scheme to extract the sediment volume using GIS 48

Figure 3-17 Adjust negative number by linear interpolation along the channel 50

Figure 3-18 Detecting negative errors 50 Figure 3-19 Modifying the errors 51 Figure 3-20 Locations are checked the sediment thickness in stream order 5 52

Figure 3-21 Bottom exposes bedrock and gravels 53

Figure 3-22 Incision to bedrock and sediment cover 53 Figure 4-1 Quaternary sediments thickness 57

Figure 4-2 Relationship between catchment area and sediment storage in Quaternary

60

Figure 4-5 Sediment accumulation rate in Holocene by catchment 6 60

Figure 4-6 Sediment buildup/time diagram 61

Figure 4-7 Denudation from mean altitude 64

Figure 4-8 Denudation from standard deviation of altitude 65

Figure 4-9 Sediment budget in the drainage basin 67

Figure 5-1 Example of calculating the EI 30 for one rain storm 74

Figure 5-2 Relation between EI30 and I60 max in the study area 75

Figure 5-3 Distribution of R-value 76

Figure 5-4 Distribution of soil types in the Onga River basin 77

Figure 5-5 Diagram defines the soil structure in Japan 78

Figure 5-6 Distribution of K-value 80

Figure 5-7 Distribution of L- value 81

Figure 5-8 Distribution of Slope value 82

Figure 5-9 Distribution of S-value 83

Figure 5-10 Rice field in steep slope with practice management 84

Figure 5-11 Distribution of CP factor 85

Figure 5-12 Soil loss map in Onga river basin 86

Figure 5-13 Soil loss map in forest area 87

Figure 5-14 A forest vs E forest by catchment with stream order 5-7 88

Figure 5-15 A forest vs E * forest by catchment with stream order 5-7 88

Figure 5-16 Specific yield-A by catchment 6 90

Figure 5-17 Specific yield-E by catchment 6 90

Figure 5-18 Relation between A-E* by catchment 5-7 91

Figure 5-19 Specific yield - A by catchment 5 92

Figure 5-20 Sediment loss - E by catchment 5 92

Figure 5-21 Relation between A-E* by catchment 5 93

Table 2-1 LS factors in USLE and RUSLE 20

Table 4-1 Sediment erosion in Onga River Basin in Holocene (8500 years BP) 65

Table 4-2 Scenarios of SDR in Onga River Basin in Holocene (8500 years BP) 68

Table 4-3 SDR in Onga River Basin in Holocene (8500 years BP) 68

Table 5-1 R-value in each year from 1990 to 2010 74

Table 5-2 Permeability code for each soil type 78

Table 5-3 K-value for each soil type 79

Table 5-4 CP-factor from some study cases in Japan 84

Table 5-5 Erosion of each catchment in forest and agriculture areas 89

Table 5-6 Evaluating the errors for catchment 5 91

CHAPTER 1

INTRODUCTION

1.1 Impacts of human activities on soil erosion and sedimentation

In globel scale, Meybeck & Vorosmarty (2005) concluded that the total river basin area directly affected by human activities is of the same order of magnitude (>

40 Mkm2) as the total area affected over the last 18000 years More than 80% of river fluxes to oceans and to internal regions have been generated in less than 30% of the

continents in natural conditions Human impacts are increasing this contrast since major new sources of material are very concentrated in intensive agriculture areas,

mining and industrial districts, and megacities Over the last 100 years, fluvial

systems have been largely impacted and modified by human activities and

anthropogenic controls, which are now to equaling the natural ones

In small scale, a research in former Soviet Union showed a clear trend of

increasing suspended sediment yield during 1949-1985, and sediment loads have

increased by about 1.4 times since that time These increased sediment loads reflect

the expansion of cultivation within the drainage basin (Bobrovitskaya, cf Walling, 1999) Dedkov and Mozzherin (1984) (cf Walling, 1999) analysis the covers

occupied by forest and cultivate to derive an approximate the magnitude of the

increasing sediment yield related to land disturbance by human activities Förster &

Wunderlich (2009) estimated the sediment stores in the catchment from the thickness

of colluvial layers and its area for each soil types, and sediment eroded from the

remained thickness of erosive layer They concluded that the sediment discharge ratio

in Holocene varies from 64.8% to 83.2% for the 301 km2 catchment, but land-use history and human impact can’t be made Macaire et al (1997) calculated the sediment budget for the Lac Chambon watershed in France for the last 15500 years

They showed that a threefold increase in erosion over the last 1400 years due to the

impact of human-induced deforestation

But, it doesn’t clear that the current erosion and deposition is over or under

‘natural value’ due to lack of long-term record of sediment transport for river in most

area of the world (Walling, 1999) Oguchi (1997) and Oguchi et al (2001) also

mention that there is less known about sediment budget on timescale of more than

1000 years in Japan, and most of researches concentrate in the central and north Japan

1.2 Long-term sedimentation, erosion and sediment discharge ratio

Natural processes such as erosion, mass movement and water cycle produce

and transport sediment from one place to another as sedimentation in natural conditions Within a catchment, it is needed to clear the relation between erosion,

deposition and discharge to manage the catchment

When human settles and disturbs the land surface without methods to protect

the soil erosion, it tends to accelerate the erosion (Syvitski et al 2005) Currently, it

is possible to measure the erosion and monitor the sediment transportation within or

discharge out of a catchment, but it is impossible to measure directly for the past

time Thus, for long-time period, sediment storage in some type of accommodations

such as reservoirs is a key which helps to know the average accumulation rate, and then, know the erosion rate of the drainage basin upper that reservoir (Einsele and

Hinderer, 1998) This erosion values in a give time, which specific for a given

environment, can be compared with the values of other time span to know how

different the environment is For examples, the Pleistocene and Holocene mechanical denudation of entire Alps derived from the fillings of perialpine lakes, demonstrated

that the average denudation rate since the last glacial maximum (17 ka B.P.) was

about 5 times higher than the Holocene rate In the Black Sea area, during the last

deglaciation (15 to 8 ka BP) the sediment yield of the rivers was two to four times greater than at present (Einsele, 2000) Another example, in Japan, Ohmori (2003)

used the sediment storage in the reservoirs to establish the denudation model

It is possible to estimate the long-term mechanical erosion rate from studies

in the landscape of particular area such as landform reconstruction This rate can be also grained from denudation-sedimentation-accumulation system of some size (lake,

larger basin) when the average accumulation rate cane be determined and the ratio of

the areas of sediment source and basin are known (Einsele, 2000) For instant,

Oguchi (1997) reconstructed five alluvial fans in Japan for Late-glacial period The sediment storage in fans was calculated from stratigraphical data Sediment yield

from each source area (fans) was estimated based on digital elevation data and the

morphometric analyses of geomorphological maps, in which all hillslopes and fluvial

surfaces were classified He estimated the supply source concluded that the ratio of sediment storage in a fan to sediment supply is between 0.3 and 0.8 The maximum

storage ratio of 0.8 reflects the ratio of washload to total load for Japanese

mountainous rivers This ratio has the same meaning with sediment delivery ratio

when the drainage basin is considered

Sediment delivery ratio is the ratio of sediment delivered at the catchment

outlet to gross erosion within the basin The magnitude of the delivery ratio for

particular basin will be influenced by a wide range of geomorphologic and

environmental factors including the nature, extent and location of the sediment sources, relief and slope characteristic, the drainage pattern and channel conditions,

vegetation cover, land use and soil structure (Walling, 1983) Understanding the

sediment delivery process at the drainage basin scale remains a challenge in erosion

and sedimentation research In the absence of reliable spatially distributed process based models for the prediction of sediment transport at the drainage basin scale,

area-specific sediment yield (SSY; tkm-2 y-1) is often assumed to decrease with increasing drainage basin area (A) However, over the last two decades several

studies reported a positive or nonlinear relation between A and SSY (de Vente et al., 2007)

In this research, an interaction between sediment erosion and deposition in

the drainage basin is considered, in order to estimate the sediment delivery ratio for

the longer time span integrated to GIS and to clarify the human activities impact on

erosions as disturbance the land-cover The Onga River basin locates in the south of Japan on Kyushu Island is chosen as study area

1.3 Objectives

For management and prediction of watershed sediment, it is important to

know its long-term sedimentation and relative factors The objective of this research are the development a systematic method integrated with GIS to estimate the natural

sedimentation and its relation to erosion through a case study in Onga River Basin in

northern Kyushu, Japan to clarify a sediment storage mechanism in the Japanese

catchment where many human activities since late Holocene were detected There are some objectives need to be done:

+ Reconstructing the Quaternary ground surface and Holocene ground surface

to estimate the sediment storage;

+ Estimating the natural sedimentation, erosion and sediment discharge of watershed and its sub-watershed for watershed management of sediment

+ Separating the natural erosion and comparing with total erosion accelerated

by human to clarify the sediment yield from human activities;

1.4 Scope

The scope of this research is determining the relation between long-term

sedimentation, erosion and discharge and estimating the impact of human activities

on the land cover in the Onga river basin for watershed management

1.5 Thesis layout

The thesis structure comprises a method for estimating the sediment volume

in the basin from the borehole data, geological map and DEM using GIS It also

includes the sediment budget calculation for the drainage basin, and showing the

relation between sheet and rill erosion from RUSLE model and total erosion estimated from denudation model An evaluation of erosion produces from

agriculture and forestry activities is clarified The thesis includes six (6) chapters and

is briefly described as:

Chapter One provides the overview of the researches on long-term and

short-term erosion and sedimentation, and on relation to pre-human impact in

comparison with human impact periods

Chapter Two introduces the study area and methods to estimate the sediment

storage, erosion for long-term and short-term

Chapter Three describes in detail how to make the palaeo-surfaces for

Holocene and Quaternary The supplement method is applied to extract the Holocene

geological boundary from DEM for modifying the geological boundary Borehole

analysis and evaluations have done to define the lower Holocene sediment layer and lower Quaternary sediment layer Among interpolation methods, the Spline method is

applied to give the suitable DEM of palaeo-surfaces

Chapter Four calculates the accumulation rates for Holocene, Quaternary

and Late Pleistocene, erosion rate from the sediment storage and total erosion in the basin A relationship between sediment storage in each catchment and its area is

established The sediment budget is applied to evaluate amount of sediment storage

in the hill slope, discharge to the channel and transport in and out of the

sub-catchment and whole catchment

Chapter Five analysis the sheet and rill erosion using RUSLE and relation to

erosion caused by disturbance of human on the land cover in forestry and agriculture

activities A relationship between parameters of RUSLE and the total erosion in Onga

River basin and its sub-catchments is proposed

Chapter Six presents the conclusions of this study

References

de By, R.A et al (2001), Principles of Geographic Information Systems, ITC, The Netherlands ISBN:90-6164-200-0

de Vente, J., Poesen, J., Arabkhedri, M and Verstraeten, G (2007), The sediment

delivery problem revisited, Progress in Physical Geography, Vol.31, p.155

Einsele, G., (2000), Sedimentary basins: evolution, facies, and sediment budget,

Springer, 2rd edition, p.454, ISBN: 3-540-66193-x

Einsele, G and Hinderer, M (1998), Quantifying denudation and sediment-accumulation systems (open and closed lakes): basic concepts and first

results, Palaeogeography, Palaeoclimatology, Palaeoecology, Vol.140, pp.7-21

Förster H., and Wunderlich J (2009), Holocene sediment budgets for upland

catchments: The problem of soilscape model and data availability, Catena, Vol.77,

pp.143-149

Hoffmann T et al (2007), Holocene floodplain sediment storage and hill slop

erosion within the Rhine catchment, The Holocene, Vol.17, No.1, pp.105-118

Macaire J.J., et al (1997), Sediment Yield during Late Glacial and Holocene Periods

in the Lac Chambon Watershed, Massif Central, France, Earth Surface Processes

and Landforms, Vol.22, pp.473-489

Meybeck M and Vorosmarty C (2005), Fluvial filtering of land-to-ocean fluxes:

From natural Holocene variations to Anthropocene, C R Geosci., Vol.337,

pp.107-123

Oguchi, T et al (2001), Fluvial geomorphology and paleo-hydrology in Japan,

Geomorphology Journal, Elsevier, Vol.39, pp.3-19

Oguchi, T (1997), Late Quaternary sediment budget in alluvialfan–source-basin

systems in Japan, Journal of Quaternary Science, Vol.12, No.5, pp.381-390

Oguchi T (1996), Late Quaternary hill-slope erosion rates in Japanese mountains

estimated from landform classification and morphometry, Zeitschrift fur

Geomorphologie Neue Folge Supplementary Band, pp.169-181

Ohmori H (2003), The paradox of Equivalence of the Davisian End-Peneplain and

Peckian primary peneplain, Concepts and Modelling in Geomorphology:

International perspective, pp.3-32

Seidel J and Mäckel R (2007), Holocene sediment budgets in two river catchments

in the Southern Upper Rhine Valley, Germany, Geomorphology, Vol.92,

pp.198-207

Syvitski J.P.M., et al (2005), Impact of Humans on the Flux of Terrestrial Sediment

to the Global Coastal Ocean, Science, Vol.38, pp.376-380

Wakamatsu, K., Matsuoka, M and Hasegawa, K (2006), GIS-based nationwide hazard zoning using the Japan engineering geomorphologic classification map,

Proceedings of the 8 th U.S National Conference on Earthquake Engineering,

San Francisco, California, USA, Paper No.849

Walling D E (1999), Linking land use, erosion and sediment yields in river basins,

Hydrobiologia, Vol.410, pp 223–240

Walling D E and Bradley S B (1988), The use of caesium-137 measurements to

investigate sediment delivery from cultivated areas in Devon, UK Sediment

Budgets (Proceedings of the Porto Alegre Symposium), IAHS Publ no.174

Walling, D.E (1983), The sediment delivery problem, Journal of Hydrology, Vol.65,

pp.209-237

CHAPTER 2

STUDY AREA AND METHODOLOGY TO ESTIMATE

EROSION AND SEDIMENTATION

2.1 Study area

Onga River located in the northern Kyushu, Japan (Figure 2-1) its drainage basin covers 1036 km2 In early-mid Holocene, the lower reach of this river acts as a lagoon in which sedimentation occurred widely to form the plain named Nogata, and

some small plains are in the upper reaches Currently, the precipitation in this area is

400 mm higher than other area in Japan, and causes many large floods (MLIT) which

are though to be one of major factors cause erosion and the main supply sources of

sediment, for example, the thickness of Holocene sediment in Nogata plain is up to

near 20 m (Shimoyama, 2002) Research on sedimentation and erosion of whole drainage basin will contribute not only for geomorphology process studies but also

for current management of the river basin and understanding the human activities

impacts.Figure 2-1 shows the location of the Onga River basin in Fukuoka prefecture and its sub-catchments on the Quaternary geological units The Nogata plain distributes from the river estuary though Onga town and Nakama city to Nogata city

Other cities include Miyawaka, Iizuka and Tagawa In this research, the river basin is

divided into 2 sub-catchments 7 and 6 sub-catchment 6 More information is

mentioned in the following sections

2.1.1 Geology and landforms

In the estuary area, Quaternary deposits include most of the late Pleistocene

sediments belong to Wakamatsu formation in the eastern part This formation

includes two members Shozugahama mud and Iwaya sand and gravel members The

Shuzugahama mud (2m thick) is exposed along the coast composed of silt with small

amount of pebble and sand (Figure 2-3) The Iwaya sand and gravel member

(15-20m) made up mainly of pebble gavel and sand with silt (Ozaki et al., 1993)

Figure 2-1 Onga River basin and its sub-catchments

Figure 2-2 Geological map of Onga River basin (modified from GSJ, 2005).

Figure 2-3 Cross section along Onga River (Shimoyama, 2002)

a: Human fill up;, b: Sanrimatsubara sand layer, c: Koyanose layer, d: Onga silt layer, e: Iwaya sand layer, f: Mid-terrace gravel, g: Heisan mud layer, h:Ancient terrace, and i:

Tertiary rock (a∼d is Holocene, e∼h is Pleistocene)

In Nogata plain, there doesn’t expose any Pleistocene sediment on the surface,

except the area around the mountain belong to upper reaches In this area there are

many tephra such as Kikai-Akahoya (K-ah), aged 6.3 ka years BP, occurs in marine deposits (Machida, 1991); AT (AIRA – Tn tephra) which is dated around 21 – 22 ka

years BP; and Aso-4 tephra (70 ka years BP) Machida and Arai (1983) shown that

the K-ah thickness of volcanic ash is about less than 20 cm, AT ash thickness is 50

cm (Nakada & Lambeck, 1998) Otherwise, there are a lot of tephra of Aso-4 flows

in the upper reaches of the study area The Aso-4 tephra interrupted when the sea

level drop between Obaradai and Misaki transgressions (Machida, 1991) and AT

formed when the sea level was lowest The pre-Quaternary rocks include Paleogene

sedimentation, Cretaceous, Permian, and Carboniferous rocks, which consist of about 70% of the basin area The K-ah ash occurs in the marine sediment, aged 6.3 ka year,

which is the geology record layer used to compare with other layers in the boreholes

in later analysis

2.1.2 Sea level change

Sea-level changes in Late Pleistocene-Holocene play a very important role to

define the ancient boundary surfaces as Holocene surface and Quaternary surface

During the Last interglacial transgression or Shimo-sueyoshi transgression, the sea level rose at peak in 130-120 ka years BP, near present sea level in the northern of

Kyushu (Kaizuka 1980) Shimoyama et al (1999) explained that this ancient sea level (marine top) in Onga, north-west of Kyushu, was -7.9 ± 1.5 m In the Last Glacial age, 20 -15 ka years BP, the sea level drop -140 to -120 m and reached the lowest peak at around 18 ka years BP in comparison with today’s one (Kaizuka 1980

& Yasuda 1990) as shown in Figure 2-4 In Late-glacial, 13 ka – 10 ka years BP, the

sea level increased to -10m lower than that of today After that the sea level withdrew

to -40 m around 11 ka to 10 ka year BP to form the coarse grain size along the seashore (Umitsu, 1991) Example, 10 m thickness pebble in Hakata bay (Karakida

et al., 1994), which underlies by fine grain size; and about 5 meter sand without

shells in Onga River mouth The sea-level rose to maximum around 6000 year BP

(Umitsu, 1991) or Jomon transgression

Figure 2-4 Relative sea level change in the Japan (Machida, 2002) from stage 5e to present

Ages in ka (1000 year) Abbreviations of volcanoes in Kyushu area: K, Kikai; At, Ata; A, Aira; Kk, Kakuto and Kirishima; Kj, kuju; Ss, Shishimuta; L, low; H, high; MIS, Marine

Isotope Stage

In the late Jomon period, around 6000 BP, the sea level higher than present 2 to

3 meter (Karakida et al., 1994) to maximum +4.0m in north of Kyushu Island (Pirazzoli, 1991), that made Onga village area become the lagoon and caused the

major deposition of the Hakata member (Karakida et al 1994.) A research in Fukue,

an island located about 200 km to the southwest of Onga River, showed that the

mean sea level is above -11.4 m at about 7.900 yr BP, and from 8000 yr BP to 5000

yr BP, mean sea level rose with an average speed of 0.40 cm/yr Then, minor

sea-level dropped, have been recognized in 5000-4000 BP and 3000-2000 BP

(Umitsu, 1991)

2.2 Method to estimate sedimentation and erosion

2.2.1 Method to estimate sedimentation

Sedimentation is a process of deposition of a solid material from a state of

suspension or solution in a fluid (usually air or water) Broadly defined it also includes deposits from glacial ice and those materials collected under the impetus of

gravity alone, as in talus deposits, or accumulations of rock debris at the base of

cliffs Estuaries or plains are filled with sediment brought in by tributaries streams,

but sedimentation in estuaries is also contributed by marine movement

Sediments can accumulate in hill-slopes, alluvial fans, river channel, flood plain, deltas and lakes bed deposits (Charlton, 2008) Colluvium is the type of deposits on or at the base of slope The sediment is discharged to the channel is call

alluvial Terraces are the storage types formed from colluvial or fluvial in the past, considered as fans For examples, many authors calculate the volume of sediment in

fans and estimate the erosion rate and deposition in Quaternary (e.g Oguchi, 1997;

Oguchi et al., 2001; de Moor and Verstraeten, 2008; Lewin et al, 2005; Harvey et al.,

1999) Sediment stored dominantly in the flood plains, and some of it distributes as bottom layers

In case of enough borehole data, the sediment volume can be calculated from two DEMs in GIS, one is the upper DEM (uDEM) and another is the lower DEM

(lDEM) The results from minus two DEMs, respectively, give the prism thickness

with the section is the cell size Total volume of sediment is the sum of all prism

thicknesses multiple with cell size (e.g., Hoffmann et al., 2007, Cohen, 2005)

cellsize lDEM

DEM u

V =∑( − )⋅ (2.1)

Figure 2-5 Example calculation of thickness from two DEMs in GIS

With the available boreholes distribute widely in the study area, this method is used to calculate the volume of sediment

2.2.2 Method to estimate erosion

Denudation is the lowering of land surface due to weathering and erosion The

average amount of this lowering in certain area over a given time span is defined as

the mean denudation rate Hereafter, for Holocene period, denudation rate is used

for mean denudation rate, and has the same meaning with erosion rate Denudation

involves the transport of disintegrated bedrock, soil and vegetation components into

continental basins or into the oceans This mass transfer is controlled by a variety of

physical, chemical, and biological processes Rainfall, surface runoff, and river flow

are the most important transport agents (Einsele, 2000) Region with high relief in conjunction with substantial rainfall are most important in delivering particular

matter into rivers

Mechanical denudation is the lowering of the landscape of a certain area by

erosion of solid material consisting of soil particles and fragments of the underlying rock Mean denudation rates indicate the mean thickness of soil or rock material

removed from the total drainage area studied for a certain period of time Therefore,

these denudation rates do not distinguish between (1) summit lowering, (2) slope

denudation, (3) rate of stream erosion in river valleys or lowlands (Einsele, 2000) In

this research, mechanical denudation is considered

The general law expresses mechanical denudation in relation to relief in

geological time period (e.g., Pinet and Souriau, 1988; Ohmori, 2003; Einsele, 2000)

The general formula can be:

D me = K dx H m (2.2)

where, D me is mechanical denudation rate, [m/Ma]; H is mean altitude, [m], m

is constant factor, and K d is a factor, varied significantly from region to region and

have to calibrate by local mechanical erosion data (Einsele, 2000)

Ohmori (2003) fit this formula (eq.2.2) to Japanese condition and proposed that the denudation rate is proportional with elevation His model bases on 82 reservoirs

throughout Japan which have sediment volume over 50,000 m3, storage capacity over 2,000,000 m3, sedimentation ratio to water storage capacity less than 25%, and observation duration longer than 10 years, with longest duration of 66 years The formula is:

β

αD

E = [mm/year] (2.3) where α and β are constants (α=4.4 10-5

; β=2.2), E is denudation rate (mm/year) and D (dispersion of altitude) is the standard deviation of altitude:

2 / 1 1

2

1

)(

n

i i

[m] (2.4)

In which H is the mean altitude given by:

n

h H

n

i i

∑

=

= 1 [m] (2.5)

or simplified as:

Most of these reservoirs used to calculate are located in the upper-most ones in

a sequence of reservoirs through individual rivers, which are considered as natural

erosion due to less affected by human activities He also proved that the model can

be applied for drainage basin (open system) and long-term period (ten thousands

years or longer) and drainage basin area larger than some tens of km2 So, this model will be used to estimate the total erosion in the basin

For a few decade or years and seasons, many physical models have been

developed and applied worldwide Most of models base on the principal that the

precipitation, surface water, slope, and human disturbance the land cover, mining disposal, etc all of these factors are contributed to accelerate the soil erosion

(Schmidt, 2000) Notable erosion models are mentioned in the next section

2.3 Erosion models

There are many erosion models established, especially notable ones such as

USLE/RULSE, WEPP, EUROSEM, etc (Schmidt, 2000)

2.3.1 Universal Soil Loss Equation

The RUSLE originated from USLE (Universal Soil Loss Equation) is an

erosion model designed to predict the short-term (few decades) average soil loss in

the disturbance of human as land-use

The USLE is an empirical equation for average annual soil loss, and is the most

commonly used to estimate the soil loss caused by overland erosion Originally, it is applied to plot scales proposed by Wischmeier and Smith (1978) An updated version

also proposed, named RUSLE Although the RULSE maintains the same structural

limitation from inherent to the USLE, the equations used to evaluate individual

factors are significantly different The greatest significant is that C-factor values cane

be estimated with the RUSLE for crops and management and tillage system where

SLRs were not available Furthermore, data gaps for estimating R-factor, the time

varying K-factor, the new algorithms for topographic factor, and the new technology

developed for estimating support practices greatly enhance RUSLE and permit its

application to modern farming practices (Lane et al.,1992) The model is:

where

A is the average annual soil loss, expressed in [txyear-1 xha-1]

R is the annual rainfall energy, expressed in [txhxMJ-1xmm-1

]

K is soil erodibility, [MJxha-1xh-1xyear-1xmm-1

]

S is the slope steepness, [non-unit]

L is the length of slope, [non-unit]

C is land cover, [non-unit]

P is practice management, [non-unit]

(1) Rainfall and runoff erosivity factor

Rainfall and runoff erosivity factor proposed by Wischmeier (1959) and

Wischmeier and Smith (1978), was derived from research data from many sources

When other factors are held constant, soil losses from cultivated fields are directly

proportional to a rainstorm parameter: the total storm energy (E) times the maximum

30-minutes intensity (I30) A rainfall factor used to estimate average annual soil loss

must include the cumulative effects of many moderate-sized storms as well as the

effect of the occasional serve ones Because of apparent cyclical patterns in rainfall

data, the rainfall records in a station should be done at least 22 years, especially in

the areas where have a large coefficient of variation of annual precipitation (Renard

et al., 1996)

A break between storms is defined as 6 h or more with less than 1.3 mm of

precipitation Rains less than 13 mm, and separated from other storms by 6 hours or

more, are omitted as insignificant unless the maximum 15 min intensity exceeds 24

mm h-1 (Wischmeier and Smith, 1978) Renard et al, (1993) proposed the modified

method to calculate the R value as follow:

Rainfall intensity for particular increment of a rainfall event (ir) is calculated

as:

r

r r

t

V i

Δ

Δ

where Δt r is the duration of the increment over which rainfall intensity is

considered to be constant, expressed in hour ΔVr is the depth of the rainfall (mm)

during the increment If ir is greater than 76 mm/h, it will be given 76 mm/h

Rainfall energy per unit depth of rainfall (er in [MJ ha-1 mm-1]) can be

1

(2.10)

To calculate the erosion index (EI) value for particular rainstorm ([MJ ha-1

mm-1 h-1], total storm kinetic energy (E) is multiplied by the maximum amount of

rain falling within 30 consecutive minutes (I30) in [mm h-1]

The average annual rainfall and runoff erosivity factor (R) [MJ ha-1 h-1 year-1

mm-1] is the average of calculated EI-values and other points where lack of data are

linear interpolation R-values in other location are interpolated by linear method

( ) ( )

j

n

j m

k

k

E n

R= ∑ ∑⎢⎣⎡ 30 ⎥⎦⎤

1

E is the total storm kinetic energy (MJ ha-1),

I 30 is the maximum 30 minute rainfall intensity (mm h-1),

j is an index of the number of year used to produce the average,

k is an index of the number of rain storms in each year,

n is the number of years used to obtain the average R,

m is number of rain storm in each year

(2) Soil erodibility factor

Soil erodibility factor (K) is related to the integrated effect of rainfall, runoff,

and infiltration on soil loss The soil erodibility is the lumped parameter that

represents an integrated average annual value of the total soil and soil profile reaction

to a large number of erosion and hydrologic processes There processes consist of soil detachment and transport by raindrop impact and surface flow, localized

deposition due to topography and tillage-induce roughness, and rain water infiltration

into the soil profile (Renard et al., 1996) This factor should be derived from soil loss

experiments, but it can be extracted from Nomograph, or using the formula:

[2 1 × 1.14× 10 4× ( 12 − ) + 2 35 × ( − 2 ) + 2 5 × ( − 3 )]× 0 1317 ÷ 100

c b

a M

where K in [t.h.MJ-1.mm-1],

M = (%silt + % sand) (100 – %clay), a = % organic matter, b = structure code,

and c = profile permeability class

K-factor is also evaluated based on geometric mean particle diameter (Dg),

express in SI unit (Romkens et al., 1996)

=

2

7101.0

659.1log

2

1exp0405.00034.0594

where Dg =exp(0.01∑ f ilnm i) [mm], with r2=0.983 (2.14) Here, f i is the primary particle size fraction in percent, and m i is the arithmetic mean of the particle size limits of that size

Eq.(2.14) is useful for predicting K values of soils for which 1) data are limited, 2) the texture composition is given in different classification system, 3) for soil of

textural extremes and well-aggregated, and 4) Only soils with less than 10% rock

fragments were considered It is certainly that the K values will less accurate than

those obtained from eq.(2.12)

(3) Slope length factor and Slope steepness factor

Slope length factor (L) and Slope steepness factor (S) are the effect of

topography on erosion These values can be calculated separately or combined

together and affect sheet and rill erosion Erosion increases as slope length increases,

and is considered by the slope length factor (L) Slope length is defined as the horizontal distance from the origin of overland flow to the point where either the

slope gradient decrease enough that position begins or runoff becomes concentrated

in a defined channel (Wischmeier and Smith, 1978) Many authors attempt to

calculate this value from DEM using GIS technology or other programming

environment (e.g., Van Remortel et al., 2001; Hickey, 2000)

Table 2-1 LS factors in USLE and RUSLE (Moore and Wilson, 1992)

LS factors*

S L = (λ/22.13)m USLE 65.4 sin2β+4.56sinβ+0.0654 m=0.5

m=0.4 m=0.3 m=0.2

tanβ>0.050.03<tanβ≤0.050.01<tanβ≤0.03tanβ≤0.01RUSLE 10.8sinβ+0.03

16.8sinβ-0.05 3sin 0.8β+0.56 (sinβ/0.0896)0.6

tanβ<0.09

tanβ≥0.09

λ≤ 4m

thawing soil with

where m=0.4, n=0.3, and A s is specific catchment area

(*) λ is slope length in [m]; β is slope angle in [degree])

Moore and Wilson (1992) proposed the equation to calculate the LS factor:

T c= (A s/22.13)m (sinβ/0.0896) n =LS (2.15)

where T c is the dimensionless sediment transport capacity, A s is the specific

catchment area, with m=0.4 to 0.6 and n=1.2 to 1.3 Griffin et al (1988) proposed

multiply with (m+1) when predicting erosion at a point

Foster and Wischmier (1974) recognized that a slope unit can not be considered

as totally uniform, so they subdivided the slope into a number of segments, which

they assumed to be uniform in slope gradient and soil properties Then,

USLE-LS-factor in this case might be calculated as

j j

m j m j j j

S LS

13.22

1

1 1 1

λλ

λλ

where L is the slope length factor for the jth segment [non-unit]

S j is the slope factor for jth segment [non-unit]

λj is the distance from the lower bounady of the j th segment to the upslope field

boundary [m]

m is the length exponent of the USLE-LS factor [non-unit]

Based on this concept, Desmet and Gover (1996) established a procedure for

automatically calculating the USLE-LS factor on topographically complex landscape

units Grid cells in DEM are considered to be segments The upslope contributing

area of each cell is defined to be considered as the contributing area (A i,j-in) at the

inler of a grid cell with coordinate (i,j) [m2] and at the cell outlet, the value at inlet

has to be increased by the grid cell area:

A i,j-out = A i,j-in +D2 (2.17)with A i,j-out is the area at the outlet of grid cell with coordinate (i,j) [m2], D is

grid cell size L-factor of USLE for grid cell with coordinates (i,j) is:

( )m j

m

m in j m

j j

x D

A D

A L

13.22

, 2

1 , 1 2 , ,

⋅

⋅

−+

+

− +

They transformed the formula proposed by Griffin et al (1988) (eq.(2.19)) to

calculate the erosion in a grid cell with coordinate (i,j) and assuming that the center

of the grid cell is representative for the whole grid cell as eq.(2.20):

=

13.22

x D

D A

13.222

21

,

2 ,

where, L i,j is the slope length factor in one cell (non-unit),

A ij-in is the contributing area at the inlet of a grid cell with coordinate (i,j), in

[m2],

x i,j is a factor with accounts for variations in the width of flow resulting from

the orientation of the cell with respect to the contour It has a value of 1.0 when the

flow exits over the side and 1.414 when the flow exits over a corner

m is the length exponent of the USLE LS-factor (non-unit),

D is the cell size, in [m],

A i,j can be considered as [Fla xD2] in GIS, substitute this value to (2.15),

j

j j

x

D Fla

m L

⋅

⋅ +

=

13 22 2

1 2

1

,

,

with Fla i,j is flow accumulation, calculated in GIS tools from DEM

The slope steepness factor reflects the influence of slope gradient on erosion

Traditional, this value is estimated in the field by use of an inclinometer, or from the

contour map It can be also estimated from DEM Many authors mentioned about this

value as mentioned in Table 2-1 S is different in USLE and RULSE

(4) Cover and Practice factors

Cover factor (C) is the cover management factor, is the ratio of soil loss from

an area with specified cover and management to that from an identical area in tilled

continuous flow It indicates how the conservation plan will affect the average annual

soil loss and how that soil loss potential will be distributed in time during

construction activities, crop rotations, or other management schemes

Practice management factor (P) is the ratio of soil loss with a support practice

like contouring, strip-cropping, or terracing to that with straight-row farming up and

down the slope

USLE/RUSLE have some advantage and limitation such as (1) the soil-loss

estimates are long-term average rates rather than precipitation-event-specific estimates, (2) there are hillslope-length and gradient limits for which the component

RUSLE equations have been verified, and (3) RUSLE does not produce

watershed-scale sediment yields and it is inappropriate to input average watershed

values for the computation of the RUSLE factors (Toy et al., 2002) However, RUSLE are frequently used for the estimation of surface erosion and sediment yield

from catchment areas from 27.9 km2 to 245 km2 (e.g., Ferro and Minacapilli, 1995;

Kothyari and Jain, 1997; Terranova, et al., 2009, Bhattarai & Dutta, 2007, Lee and

Lee, 2006, etc.)

2.3.2 Water Erosion Prediction Project model

Water Erosion Prediction Project (WEPP) model is a daily time-step simulation model which uses the rill-interrill concept of soil erosion (Foster, 1982) Computes

soil loss along a slope and sediment yield at the end of a hillslope and is a daily

simulation model The appropriate scales for application are tens of meters for

hillslope profiles, and up to hundreds of meters for small watersheds For scales greater than 100 meters, a watershed representation is necessary to prevent erosion

predictions from becoming excessively large (Flanagan et al 1995) This model has

been developing by USDA aimed at replacing the USLE in 1995 WEPP is to be

delivered in three versions: profile, watershed and grid (Laflen et al., 1997)

Figure 2-6 Schematic of a small watershed which the WEPP erosion model could be

applied to Individual hillslopes (1 to 5), or the entire watershed (composed of 5 hillslopes, 2 channel segments, and 3 impoundments) could be simulated (Flanagan et al., 1995)

The minimum required data types are climate (precipitation, temperature, solar

radiation, wind information), slope (slope orientation, slope length, and slope

steepness), soil (physical and hydrological parameters), cropping/management (tillage sequences and implement parameters, plant and residual management, initial

conditions, contouring, subsurface drainage, crop rotation), irrigation (irrigation is

used on channels), but for watershed simulation, the additional data types are hill slope information pass (all information from each hill slope), structure (watershed

configuration), channel (length, with and slope) and impoundment (if yes)

The advantages of this model include capabilities for estimating spatial and

temporal distributions of soil loss, and since the model is process-based it can be extrapolated to a broad range of conditions that may not be practical or economical

to field test But, the appropriate scales for application are tens of meters for hillslope

profiles, and up to hundreds of meters for small watersheds For scales greater than

100 meters, a watershed representation is necessary to prevent erosion predictions from becoming excessively large In addition, the model requires a lot type of input

data, especially detail data, and is the daily simulation model It is out of purpose of

this research

USLE/RUSLE has been developed historically for along time, and applied widely around the world from plot scale, small catchment to large catchment This

model predict the soil loss estimates for long-term (few decades) average rate rather

than precipitation-event-specific estimate as WEPP or EUROSEM There are some

application to estimate surface erosion and sediment yield from catchment areas 27.9

km2 to 245 km2

2.4 Unit conversion

Unit conversion is used to convert from denudation expressed in [mm/year] to

[ton/ha/year], storage [m3/km2/year] to [ton/ha/year], and vice versa Einsele (2000)

proposed a formula to convert from denudation rate to storage and vice versa (Figure

2-7)

Figure 2-7 Units conversion graph (Einsele, 2000)

For example, DR=1, ρ=2.6, n=0.4,

SY=DR/F D=2*102/3.85=5.2t/km2/a; V s =SY*F s=5.2*0.63=3.3m3/km2/a

And another, V s=25, n=0.6, ρ=2.4,

SY=V s /F s=25/0.94=27t/km2/a; DR=SY*FD=27*4.17*10-2=1.1cm/ka

Average density of rock is 2.7 t/m3, dry bulk density of silt layer is 1.0

2.5 Conclusions

In early-mid Holocene, the lower reach of Onga River was filled by a lot of

sediment to reach the thickness up to about 20 m On other hand, heavy rains and

many large floods have been occurred prior the water project establishment, which are advantages for erosion and transporting the sediment upper reaches to lower

reaches

The sea level change in Late Quaternary effected strongly to the sedimentation

and erosion in the lower reach of the basin Beside, the tephras, which is important for dating the sediment, and shell mounds occurred widely in the study area,

especially in the marine sediment Thus, help to define the relative ages of sediment

contained them It is a good environment to research the sedimentation and erosion

of the basin for a long and short period of times

To estimate the sedimentation and erosion, some methods are proposed

Among many short-term erosion models, USLE/RUSLE is widespread applied not

only for plot scale but also for catchment scale from 27.9 km2 to 245 km2, although there are some cautions as applying for catchment and outside the validated area To

overcome the scale limitation, the catchment should be divided into smaller ones

Some researches adapt the parameters of USLE, and apply that model in Japanese

condition (e.g Taneda, 1980 and 1981; Kitahara et al., 2000; Shiono et al., 2002; Yoshikawa et al., 2004) For long-term periods, the denudation model proposed by

Ohmori is suitable to predict the total erosion the basin

In supporting by GIS and a wide distribution of borehole data, the volume

method can be applied to estimate the sediment storage

References

Arnold, J G., Weltz, M A , Alberts, E E and Flanagan, D.C (1995), Plant growth component, USDA-Water Erosion Prediction Project Hillslope Profile and Watershed Model Documentation, pp.8.1-8.41

Bhattarai, R & Dutta, D (2007), Estimation of Soil Erosion and Sediment Yield Using GIS at Catchment Scale Water Resour Manage, Vol.21, pp.1635–1647,

DOI 10.1007/s11269-006-9118-z

Charlton, R (2008), Fundamentals of fluvial geomorphology ISBN: 0-203-371089 Cohen, K.M (2005), 3d geostatistical interpolation and geological interpretation of paleo–groundwater rise in the Holocene coastal prism in the Netherlands, River Deltas—Concepts, Models, and Examples, No.83, pp.341-436

De Moor, J.J.W and Verstraeten, G (2008), Alluvial and colluvial sediment storage

in the Geul River catchment (The Netherlands) — combining field and modelling data to construct a Late Holocene sediment budget, Geomorphology, Vol.95, pp

Einsele, G., (2000), Sedimentary basins: evolution, facies, and sediment budget,

Springer, 2rd edition, p.454, ISBN: 3-540-66193-x

Ferro, V., Minacapilli, M (1995), Sediment delivery processes at basin scale, Hydrol Sci J, Vol.40, No.6, pp.703–717

Flanagan, D.C., Ascough II, J.C., Nicks, A.D., Nearing, M.A and Laflen, J.M., (1995), Overview of the wepp erosion prediction model, USDA-Water Erosion Prediction Project Hillslope Profile and Watershed Model Documentation,

Gaillard, M.J., Dearing, J.A., Daoushy, F El., Enell, M and Hatkansson, H (1991),

A late Holocene record of land-use history, soil erosion, lake trophy and lake-level fluctuations at Bjaresjosjon (South Sweden), Journal of Paleolimnology, Vol.6, pp.51-81

Geological Survey of Japan (GSJ), (2005) Digital Geoscience Map Series G20-7, Digital Geological Maps of Japan 1:200,000, Western part of Chugoku, Kyushu and Nansei Shoto, http://www.gsj.jp/Map/index_e.html

Gray, D.M (1961), Interrelationships of Watershed Characteristicsm, J Geophysical Research, Vol.66, No.4, pp.1215-1223

Griffin, M.L., Beasley, D.B., Fletcher, J.J., and Foster, G.R (1988), Estimating soil loss on topographically non-uniform field and farm units, J Soil and Water Cons.,

Vol.43, pp.326-331,

Harvey, A.M., Wigand, P.E and Wells, S.G (1999), Response of alluvial fan systems

to the late Pleistocene to Holocene climatic transition: contrasts between the margins of pluvial Lakes Lahontan and Mojave, Nevada and California, USA,

Kobayashi, Y., Kitahara, H., and Ono, H (2004), Surface erosion from landslide site

in weathered granite area and the analysis by using USLE, J.Jpn For.Soc Vol.86,

pp.365-371