Morphological Changes of the Lower Tedori River, Japan, Over 50 Years

Home , Degradation (geology), River morphology, River mouth

Article Morphological Changes of the Lower Tedori River, Japan, over 50 Years

Dang Minh Hai 1,*, Shinya Umeda 2 and Masatoshi Yuhi 2 1 Faculty of Water Resources Engineering, Thuyloi University, 175 Tay Son, Dong Da, Hanoi 116705, Vietnam 2 Faculty of Geosciences and Civil Engineering, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan * Correspondence: [email protected]; Tel.: +84-38-699-7676

Received: 30 July 2019; Accepted: 3 September 2019; Published: 5 September 2019

Abstract: Long-term variation in the morphological characteristics of rivers is complicated as a result of temporally and spatially variable natural processes and anthropogenic disturbances. Better understanding of their relationship is therefore important for river basin management. The present study conducted a detailed analysis of a long-term dataset consisting of a 58-year topographic survey and related data on human impact to clarify the long-term variation in the morphological characteristics of the lower Tedori River in Japan. An empirical model was established through the assessment of spatiotemporal variations in nonlinear rates of adjustment. The results indicated that sediment extraction and dam construction profoundly aﬀected the morphological characteristics of the lower Tedori River and that the vertical adjustment of the river channel experienced ﬁve phases. Degradation occurred from 7 to 16 km upstream from the river mouth over all phases. Two aggradation phases following two degradation phases were observed from the river mouth to 2 km upstream. Aggradation and degradation phases appeared alternately from 2 to 7 km. The representative nonlinear rates of vertical adjustment in the second phase were the highest compared with those in the other phases in the entire reach. The correlation analysis revealed that the incision phase was mostly coupled with channel narrowing, while widening followed the deposition phase. It was deduced from aerial photo analysis and a comparison between slope and empirical critical slope that the pattern in the lower Tedori River was braided during the period 1950–2000.

Keywords: Tedori River; degradation; sand and gravel mining; dam construction

1. Introduction The long-term variation in the morphological characteristics of rivers is complicated as a result of temporally and spatially variable natural processes and anthropogenic disturbances. Natural processes can progress exceedingly slowly and be practically imperceptible by human standards [1]. In contrast, human activities can have a significant and rapid impact on natural processes and trends, resulting in a compressed time scale for river adjustments [2]. Many researchers have made efforts to equate the morphological responses of rivers to associated driving factors. In order to sustainably manage river systems, it is necessary to further investigate the characteristics of variation in river morphology at various temporal and spatial scales. In Japan, the morphology of pristine rivers is due to the combination of steep landforms and frequent storms, and changes are often accelerated by the effects of earthquakes and volcanic eruptions [3]. Japanese rivers differ from most continental rivers in the world because they are short, steep, and exhibit very flashy flow regimes [4]. Since around 1960, when Japan entered a period of high economic growth, a substantial amount of sand and gravel has been extracted from rivers to be used as concrete aggregate. Additionally, many dams have been built to control floods, generate electricity,

Water 2019, 11, 1852; doi:10.3390/w11091852 www.mdpi.com/journal/water WaterWater2019 2019, 11, ,11 1852, x FOR PEER REVIEW 2 of2 17 of 17

electricity, and secure water for various uses. This has seriously affected the morphology of several andJapanese secure waterrivers. for Several various previous uses. This studies has seriouslyon Japanese aff ectedrivers the clarified morphology the long of-term several impacts Japanese of sand rivers. Severaland gravel previous mining studies and/or on Japanese dam construction rivers clarified on riverbed the long-term degradation. impacts The of sand correlation and gravel between mining andvariatio/or damn in construction the riverbed on elevation riverbed, sand degradation. and gravel The extraction correlation, and betweendam construction variation was in the observed riverbed elevation,in the lower sand Tenryu and gravel River extraction, [5]. Sato andet al. dam [6] indicated construction thatwas dueobserved to the dredging in the lower of a sandbar Tenryu Riverin the [ 5]. Satoearly et al.1970s, [6] indicated the longitudinal that due profile to the of dredging the Samegawa of a sandbar River in the region early 1970s, 1 km from the longitudinal the river mouth profile ofdegraded the Samegawa significantly. River in the region 1 km from the river mouth degraded significantly. Recently,Recently, the the Tedori Tedori River, River, located located in the in mid-northern the mid-northern part of part Japan of Japan (Figure (Figure1), has 1), been has seriously been affseriouslyected by affected continuous by continuous anthropogenic anthropogenic activities activities such as such dredging, as dredging, sand sand and and gravel gravel mining, mining, and multipurposeand multipurpose construction. construction. These These activities activities have have disturbed disturbed the previous the previous equilibrium equilibrium condition condition of the river.of the In particular,river. In particular, a significant a significant amount amount of sediment of sedim wasent moved was moved out due out to due mining to mining and was and trapped was intrapped the Tedorigawa in the Tedorigawa Dam. This Dam. induced This intensiveinduced intensive changes changes in the morphological in the morphological characteristics characteristics between 1950between and 2007. 1950 Dangand 2007. et al. Dang [7,8 ]et conducted al. [7,8] conducted a detailed a detailed analysis analysis of temporal of temporal and spatial and spatial trends trends in the verticalin the adjustmentvertical adjustment associated associated with human with human activities activities at the at lower the lower Tedori Tedori River River from from 1950 1950 to 2007.to 2007. However, a mutual relationship among hydromorphological characteristics of the Tedori River However, a mutual relationship among hydromorphological characteristics of the Tedori River has not has not been fully addressed yet. been fully addressed yet.

FigureFigure 1. Study area. area. This paper aims to quantify variations in several morphological characteristics of the lower Tedori This paper aims to quantify variations in several morphological characteristics of the lower River.Tedori First, River. magnitude First, magnitude and nonlinear and ratesnonlinear of vertical rates of adjustment vertical adjustment are captured. are Then, captured. variations Then, in bankfullvariations width, in slope, bankfull specific width, stream slope, energy, specific and stream mean energy,sediment and diameter mean are sediment described. diameter Finally, are the intercorrelationdescribed. Finally, among the morphologicintercorrelation characteristics among morphologic is discussed. characteristics The paper is discussed. adds novel The findings paper to prioradds studies novel dealingfindings with to prior similar studies Japanese dealing rivers, with suggesting similar Japanese an empirical rivers, model suggesting of riverbed an empirical elevation changemodel using of riverbed an intercorrelated elevation analysis change of using river an morphology intercorrelated characteristics. analysis of river morphology characteristics. 2. Study Area 2. Study Area 2.1. General Setting 2.1.The General Tedori Setting River, which originates at Mt. Hakusan, with the highest elevation of 2702 m, has a 2 catchmentThe areaTedori of 809River, km whichand originates a channel at length Mt. Hakusan, of 72 km with (Figure the1 highest). The main elevation tributary of 2702 of them, has Tedori a Rivercatchment is the Ushikubi area of 809 River, km2 and with a otherchannel tributaries length of being72 km the(Figure Ozo 1). River, The main the Dainichi tributary River, of the and Tedori other smallRiver rivers. is the Ushikubi The lower River, Tedori with River other (LTR)tributaries referred being to the in Ozo the River, remainder the Dainichi of this River, paper and (the other reach downstreamsmall rivers. of The the lowerTedorigawa Tedori Dam) River is (LTR) the portion referred between to in the the remainder river mouth of this and paper 16 km (the upstream. reach The Tedori River is one of the steepest in Japan, with an average slope of 0.037 and 0.0069 for the entire stretch of the river and the LTR, respectively. The river pattern of the LTR is mainly braided. Based on Water 2019, 11, 1852 3 of 17 homogeneity related to slope, width, and the mean diameter of bed material, the LTR can be divided into four subreaches (SRs) located the following distances from the river mouth: SR 1: 0–2 km: SR 2: 2–7 km; SR 3: 7–13 km; and SR 4: 13–16 km [9]. The climate of the catchment area is dominated by monsoon winds blowing in from the Japan Sea. The mean annual rainfall in the Tedori River catchment is about 2600 mm/yr on the plains and 3300–3600 mm/yr in the mountain ranges. The drainage area is underlaid by various lithologies, including ancient Hida metamorphic rocks and volcanic rocks from eruptions of Mt. Hakusan. Nobi rhyolites (pyroclastic rock) from the Mesozoic to the Cenozoic eras are distributed around the Mt. Hakusan area, while Tedori layers are present in the area east of the Tedorigawa Dam and the Ozo River. Both the Tedori layers and Nobi rhyolites are prone to massive collapse. Therefore, a large amount of gravel- and sand-sized sediment has been flowing into the Tedori River.

2.2. Human Impact on the Tedori River Basin Recently, several anthropogenic activities have exerted significant influence on the behavior of the LTR to different extents. Such activities, including debris dam construction, dredging, sand and gravel mining, and dam construction have reduced the sediment supply and changed the flow regime of the LTR. Since the 1910s, approximately 150 debris dams have been erected to monitor and control debris production and sediment runoff along the upstream segment of the river in the Hakusan Mountains. Sediment trapped by debris dams amounted to 9.2 106 m3 by 1991. To enhance the conveying × capacity of the LTR, dredging activity (DA) was performed in the 0–4.8 km subreach from 1949 to 1963. The total volume of dredged sediment was 2.1 106 m3. × In the post-World War II period, there was intensive sand and gravel mining (SGM) in the area from the river mouth to 15 km upstream. During the mid-1960s, SGM was at its highest rate, causing substantial degradation of the riverbed. To deal with this problem, a law was passed in the mid-1970s to prevent the overmining of sand and gravel. Mining was then completely prohibited in 1991. The total amount of sand and gravel extracted was 8.9 106 m3, computed on the basis of licenses issued by × local authorities over the period 1950–1991. In reality, the volume of mined sediment could be greater than the official volume [10]. Moreover, due to Tedorigawa Dam operation (TDO) from 1980 and several other dams in prior years (Figure1), all the bed load and some of the suspended sediment originating from one of the main tributaries (the Ushikubi River) were trapped. The accumulated sediment retained by the Tedorigawa Dam amounted to 8.2 106 m3 by 2005. The aforementioned human activities significantly reduced × the bed sediment volume of the downstream Tedori River, consequently inducing a fatal imbalance between supplied and transported sediment in this area. Furthermore, the operation of a number of dams, especially TDO from 1980, has played an important role in altering the flow regime and sediment transporting capacity. Based on flow data measured at the Tsurugi station about 14 km upstream from the river mouth (Figure1), two separate hydrologic periods were considered, 1928–1979 and 1980–2006. Two series of data indicated that TDO reduced the magnitude and frequency of large floods. The mean annual maximum flood discharge (1144 m3/s; recurrence interval (RI) = 2.5 years) in the post-dam period was 35% less than in the pre-dam period (1755 m3/s; RI = 2.3 years). In this study, these values were taken as reference channel-forming discharges.

3. Materials and Methods

3.1. Data Datasets were provided by the Hokuriku Regional Development Bureau of the Ministry of Land, Infrastructure, Transport, and Tourism of Japan (HRDB). Thirty-six annual surveys of the riverbed topography (1950–1979, 1991, 1997, 1998, 2002, 2003, and 2007) were collected. Each annual survey consisted of 81 river cross-sections at intervals of 200 m along the 0–16 km reach. The bed elevation Water 2019, 11, x FOR PEER REVIEW 4 of 17 Water 2019, 11, 1852 4 of 17 consisted of 81 river cross-sections at intervals of 200 m along the 0–16 km reach. The bed elevation was measured relative to the Tokyo Peil (TP), which is the standard ground elevation in Japan based on the mea meann sea level level of Tokyo Bay. Then, volumes of sand and gravel mined from the river channel were gathered gathered from from 1951 1951 to to 1991 1991 on on the the basis basis of annual of annual permissions permissions issued issued by local by local governments. governments. The volumeThe volume of material of material dredged dredged annually annually in maintenance in maintenance activities activities was ob wastained obtained for the for period the period1949– 1963 as well. Moreover, the annual maximum river discharge (Qmax) measured at the Tsurugi station 1949–1963 as well. Moreover, the annual maximum river discharge (Qmax) measured at the Tsurugi wasstation examined was examined from 1928 from to 19282006. to Finally, 2006. Finally, sequential sequential aerial photographs aerial photographs at a scale at a of scale 1:15,000 of 1:15,000 were obtainedwere obtained for the for years the years 1955, 1955, 1968, 1968, 1984, 1984, 1995, 1995, and and 2000. 2000. In In general, general, the the 58 58-year-year span span of of the the data surveyed with with high high resolution resolution provided provided good good data data coverage coverage in in time time and and space space to to capture the main trendstrends of long long-term-term variation in the river morphology and its controlling factors.factors.

3.2. Method Methodss

3.2.1. Vertical Vertical Adjustment Rate The characteristics of the vertical adjustment of of the riverbed are best mathematically expressed by nonlinear decay functions [1]. [1]. However, However, the mathematical form of the function strongly depends on the specific specific characteris characteristicstics of of the the river. river. Based Based on on the the availability availability of of cross cross-sectional-sectional data data of of the the LTR, LTR, an empirical empirical model of variation of riverbed riverbed elevation was established established for two periods, 1950 1950–1979–1979 and 19911991–2007.–2007. Note Note that that there there is a gap between the two periods because of a l lackack of field field data. Through observing severalseveral representativerepresentative cross-sections, cross-sections, we we recognized recognized that that riverbed riverbed elevation elevation experienced experienced five fivephases phases of vertical of vertical adjustment adjustment (Figure (Figure2); the 2) first; the threefirst three phases phases from 1950–1979,from 1950– the1979, fourth the fourth phase phase from from1980–1991 1980– (gap1991 period),(gap period), and the and fifth the phase fifth phase from 1991 from to 1991 2007. to It 2007. is known It is known that among that among various various human humanactivities, activities, bed sediment bed sediment extraction extraction causes instant causes degradation instant degradation of the riverbed of the elevation. riverbed During elevation. the Duringperiod 1950–1979,the period SGM1950– was1979, conducted SGM was in conducted the reach ofin 0–15the reach km. Therefore,of 0–15 km. depending Therefore, on depending the sediment on thevolume sediment extracted, volume variation extracted, in the variation riverbed in occurredthe riverbed in three occurred phases in duringthree phases 1950–1979. during In 1950 the– fourth1979. Inphase, the fourth variation phase, in the variation riverbed in wasthe riverbed dominated was by dominated both SGM by and both TDO. SGM The and fifth TDO. phase The was fifth aff phaseected wasonly affected by TDO. only by TDO.

FigureFigure 2. FittingFitting exponential exponential equations equations to to temporal temporal vertical vertical adjustment adjustment at the 62nd cross cross-section.-section.

To clarify temporally and spatially nonlinear variations in riverbed elevation, the nonlinear functionfunction should bebe usedused toto fitfit thethe elevation elevation measured measured over over time time at at a a site site [1 [1,2],2]. Temporal. Temporal adjustment adjustment of ofriverbed riverbed elevation elevation at a at cross-section a cross-section of the of Tedori the Tedori River can River be described can be described well by exponential well by exponential functions. functions.The general The form general is: form is: bt Z(t) (Zmin c) = ae (1) Z(t)− − (Zmin − c) = aebt (1) where Z(t) is riverbed elevation averaged at a cross-section for a given year, in TP; Zmin is the lowest where Z(t) is riverbed elevation averaged at a cross-section for a given year, in TP; Zmin is the lowest riverbed elevation during a given period for each cross-section, in TP; a is the coefficient determined by riverbed elevation during a given period for each cross-section, in TP; a is the coefficient determined regression analysis, representing riverbed elevation at initial year to = 0 deduced from the regression by regression analysis, representing riverbed elevation at initial year to = 0 deduced from the model, in TP; b is the coefficient determined by regression analysis, expressing the rate of change in regression model, in TP; b is the coefficient determined by regression analysis, expressing the rate of

Water 2019, 11, 1852 5 of 17 riverbed elevation per year: b > 0 for degradation, b < 0 for aggradation; c is the coefficient ensuring that the left side of Equation (1) is always positive, and c was taken as 1 m for this study; and t is the time since the onset of each phase in the adjustment process, in years. To eliminate the effect of the longitudinal position of cross-sections on coefficient b, riverbed elevation was extracted by (Z c) before being fitted by the exponential function as shown in min − Equation (1). Equation (1) was then applied to all cross-sections. Note that transitional years between two successive phases were addressed based on the criterion that goodness of fit with the riverbed was simultaneously attained in both phases. Accordingly, the transitional years between phases I and II varied from 1958 to 1962 and between phases II and III from 1968 to 1973, depending on the location of the cross-section. The vertical adjustment rate (b-value) for phases I, II, III, and V was determined from Equation (1). For phase IV, due to limited data, b-value was calculated by dividing the vertical adjustment by the duration of the fourth phase.

3.2.2. Bankfull Width To investigate the relationship between horizontal and vertical adjustment of the river channel, temporal changes in bankfull channel width were examined. Bankfull channel width at each cross-section was defined based on the width of the water surface at bankfull discharge. Leopold [11] stated that most investigations have concluded that the recurrence intervals of bankfull discharge ranged from 1.0 to 2.5 years. Here, bankfull discharge was taken to be equal to mean annual maximum flood discharge in the pre-dam period (1755 m3/s; RI = 2.3 years). Bankfull channel width was then estimated based on numerical simulations using the RIC–Nays model [12]. The RIC–Nays model is a two-dimensional process-based numerical model for flow and morphology. Two-dimensional unsteady flow is computed based on the shallow-water equations expressed in a general coordinate system. The morphological computation involves a combination of flow field, sediment transport, and bed level changes. The hydraulic model established for the LTR was first calibrated and validated against the water marks measured at the historic flood events in 2002 and 2006. The model was then simulated with a bankfull discharge of 1755 m3/s and the riverbed topographies of 1950, 1960, 1970, 1978, 1991, and 2002, and the bankfull channel width of each cross-section was computed for those years.

3.2.3. Total Stream Power and Specific Stream Power Total stream power (TSP; Wm–1) is defined as the total supply of kinetic power per unit length of the stream. TSP is mostly dissipated in maintaining fluid flow against fluid shear resistance [13]:

TSP = γQS (2)

The specific stream power (SSP; Wm–2) quantifies the potential energy dissipation of a column of fluid over unit bed area: TSP SSP = (3) W where γ is the unit weight of water (9800 N/m3); Q is discharge (m3/s), which is the mean annual flood discharge, 1144 m3/s for the post-dam period (1980–2007) and 1755 m3/s for the pre-dam period (1950–1979); S is energy slope (m/m), which is nearly equal to bed slope; and W is the bankfull channel width (m).

4. Results and Discussion

4.1. Vertical Variation in Riverbed Elevation The characteristics of vertical adjustment significantly change at various temporal and spatial scales. In order to examine the variability, annual longitudinal profiles were first established based on Water 2019, 11, 1852 6 of 17 the mean riverbed elevation at each cross-section. The vertical adjustment of each cross-section over a period was then computed as the change in mean riverbed elevation between the first and last year Water 2019, 11, x FOR PEER REVIEW 6 of 17 of each period. Furthermore, the vertical adjustment of a subreach (∆zs) was defined by averaging the vertical adjustments of cross-sections located in each SR. Figure3a shows that an overall eroding averaging the vertical adjustments of cross-sections located in each SR. Figure 3a shows that an tendencyoverall was eroding observed tendency during was the observed period during 1950–2007 the period in the LTR.1950– The2007 degradationin the LTR. The was degradation highest at SR 3, with ∆zs of 2.74 m, and lowest at SR 1, with ∆zs of 0.77 m. The amount of degradation of SR 2 and was highest− at SR 3, with ∆zs of −2.74 m, and lowest− at SR 1, with ∆zs of −0.77 m. The amount of SR 4degradation was nearly equal,of SR 2 withand SR∆ z4s wasof nearly2.06 m equal, and with2.2 m,∆zs respectively.of −2.06 m and −2.2 m, respectively. − −

(a) (b)

(e) (f)

FigureFigure 3. Vertical 3. Vertical adjustment adjustment of riverbed of riverbed elevation: elevation: (a ()a vertical) vertical adjustment adjustment inin thethe periodperiod fromfrom 1950–2007,1950– (b) vertical2007, (b adjustment) vertical adjustment in phase in I, phase (c) vertical I, (c) vertical adjustment adjustment in phase in phase II, (d II,) vertical(d) vertical adjustment adjustment in in phase III, (ephase) vertical III, (e adjustment) vertical adjustment in phase in IV, phase and IV, (f) and vertical (f) vertical adjustment adjustment in phase in phase V. V.

4.2. Variation4.2. Variation in b-Value in b-Value In orderIn order to quantify to quantify the the deviation deviation and and medianmedian of of nonlinear nonlinear rates rates of ofvertical vertical adjustment adjustment (b-value) (b-value) at four SRs for different phases, we used box plots (Figure 4). The box plots indicate lower quartile, at four SRs for diﬀerent phases, we used box plots (Figure4). The box plots indicate lower quartile, median, upper quartile, and interquartile range (IQR) of b-value. Whiskers from each end of the box b median,show upper the data quartile, range or, and if interquartilesmaller, the most range extreme (IQR) values of -value. within Whiskers1.5 times the from interquartile each end range of the box

Water 2019, 11, 1852 7 of 17 show the data range or, if smaller, the most extreme values within 1.5 times the interquartile range from theWater ends 2019, of11, thex FOR box. PEER Outliers REVIEW beyond the ends of the whiskers are marked with “o”. Here,7 of 17 the median of a box plot is considered to represent the nonlinear rate of vertical adjustment of an SR, and IQRfrom (Figure the end5)s as of athe measure box. Outliers of the beyond strength the of ends deviation of the whiskers of vertical are adjustmentmarked with rates“o”. Here, in each the SR. In phasemedian I, homogeneous of a box plot is distribution considered to with represent a relatively the nonlinear narrow rate range of vertical was observedadjustment at of all an subreaches SR, and (FigureIQR4). Nearly(Figure 5) all asb a-values measure were of the negative, strength exceptof deviation for some of vertic dataal greateradjustment than rates the in ﬁrst each quartile SR. In at phase I, homogeneous distribution with a relatively narrow range was observed at all subreaches SR 2 and an outlier in SR 3. The representative vertical adjustment rates hardly changed in space. (Figure 4). Nearly all b-values were negative, except for some data greater than the first quartile at SR The variation was in the range from 0.02 to 0.01 yr 1. Spatial variation in b-value at SR 3 was less 2 and an outlier in SR 3. The representative− − vertical− adjustment rates hardly changed in space. The than thatvariation at other was subreaches. in the range The from representative −0.02 to −0.01 yr erosion−1. Spatial rates variation of four in subreaches b-value at SR in 3 phase was less II were than the highestthat compared at other subreaches. with other The phases. representative SR 4 showed erosion symmetric rates of four distribution, subreaches whereas in phase theII were remaining the subreacheshighest had compared skewed with distribution. other phases. The SR lowest 4 showed representative symmetric verticaldistribution, adjustment whereas rate the wasremaining found at SR 2. IQRsubreaches for SR had 1, SR skewed 2, and distribution. SR 3 became The large lowest in representative phase II compared vertical with adjustment other phases rate was (Figure found 5at). 1 InSR phase 2. IQR III, for the SR highest 1, SR 2, depositionand SR 3 became rate inlarge the in range phase of II 0.03–0.06compared yrwith− occurredother phases at the(Figure whole 5). SR 1, whereas a seriousIn phase erosion III, the highest rate from deposition0.09 to rate0.04 in the yr range–1 was of observed0.03–0.06 yr at−1 almost occurred the at entirethe whole range SR of1, SR − − –1 3. Bothwhereas deposition a serious and erosion erosion rate were from observed −0.09 to −0.04 at SR yr 2 andwas observed SR 4. The at rangealmost ofthe deviation entire range in of vertical SR 3. Both deposition and erosion were observed at SR 2 and SR 4. The range of deviation in vertical adjustment rates was similar at SR 3 and SR 4 (Figure4). In phase III, IQR of SR 1 was the minimum, adjustment rates was similar at SR 3 and SR 4 (Figure 4). In phase III, IQR of SR 1 was the minimum, while IQR of SR 2 was the maximum. In phase IV, the representative rate and IQR of SR 2 were the while IQR of SR 2 was the maximum. In phase IV, the representative rate and IQR of SR 2 were the same assame those as those of SR of 3. SR Although 3. Although the the median median vertical vertical adjustment adjustment rate of of SR SR 1 1was was nearly nearly equivalent equivalent to to that ofthat SR 4,of theSR 4, range the range of deviation of deviation in the in the vertical vertical adjustment adjustment rate rate of of SR SR 11 waswas nearlynearly half half that that of of SR SR 4. In phase4. In V, phase there V, was there a gradual was a gradual decrease decrease in representative in representative vertical vertical adjustment adjustment rates rates from from SRSR 1 to SR 4. A depositionSR 4. A deposition trend was trend exhibited was exhibited at SR at 1 SR and 1 and SR SR 2, whereas2, whereas erosion erosion occurredoccurred at at SR SR 4. 4.SR SR3 was 3 was consideredconsidered as transitional as transitional because because both both deposition deposition and and erosion erosion werewere witnessed witnessed there there (Figure (Figure 4).4 The). The rangesranges of deviation of deviation in vertical in vertical adjustment adjustment rates rates for for SR SR 1, 1, SR SR2, 2, andand SR 4 4 were were similar similar and and smaller smaller than than that ofthat SR 3of (Figure SR 3 (Figure5). 5).

(a) (b)

Figure 4. Cont.

WaterWaterWater2019 2019 2019, ,11 11, ,11, 1852 x, xFOR FOR PEER PEER REVIEW REVIEW 88of 8of of 1717 17

(e()e ) Figure 4. Temporal and spatial variations in nonlinear rate of vertical adjustment (b-value). FigureFigure 4. 4.Temporal Temporal and and spatial spatial variations variations in in nonlinear nonlinear rate rate of of vertical vertical adjustment adjustment (b -value).(b-value).

Figure 5. Temporal and spatial variations in interquartile range of the vertical adjustment rate. FigureFigure 5. 5. Temporal Temporal and and spatial spatial variations variations in in interquartile interquartile range range of of the the vertical vertical adjustment adjustment rate. rate. Overall, b-values extracted from Equation (1) fully captured the temporal and spatial trends in Overall,Overall, b -bvalues-values extracted extracted from from Eq Equationuation (1) (1) fully fully captured captured the the temporal temporal and and spatial spatial trends trends in in verticalvertical adjustment.adjustment. TheThe variationvariation ofof trendstrends inin representativerepresentative bb-values-values waswas analagousanalagous toto thatthat inin∆ ∆zzs s vertical adjustment. The variation of trends in representative b-values was analagous to that in ∆zs mentioned previously. This proves that the way to define b-values based on Equation (1) could be mentionedmentioned previously. previously. This This proves proves that that the the way way to to define define b -bvalues-values based based on on Eq Equationuation (1) (1) could could be be better than methods used in several previous studies. Rinaldi and Simon [2] utilized a dimensionless betterbetter than than methods methods used used in in several several previous previous studies. studies. Rinaldi Rinaldi and and Simon Simon [2] [2] utilized utilized a adimensionless dimensionless exponential function to extract a nonlinear rate of vertical adjustment (k) for riverbed elevation in exponentialexponential function function to to extract extract a anonlinear nonlinear rate rate of of vertical vertical adjustment adjustment (k ()k for) for riverbed riverbed elevation elevation in in the the the Arno River, central Italy. In their study, the resulting variations in k could not totally capture the ArnoArno River, River, central central Italy. Italy. In In their their study, study, the the resulting resulting variations variations in in k kcould could not not totally totally capture capture the the characteristics of vertical adjustment of the Arno River. For example, it was seen that the values of k in characteristicscharacteristics of of vertical vertical adjustment adjustment of of the the Arno Arno River. River. For For example, example, it it was was seen seen that that the the values values of of k k different phases were nearly the same, while the corresponding degradation amounts of these phases inin different different phas phaseses were were nearly nearly the the same, same, while while the the corresponding corresponding degradation degradation amounts amounts of of these these were quite different. Dang et al. [7] applied an exponential function to define the nonlinear rate of phasesphases were were quite quite different. different. Dang Dang et et al. al. [7] [7] applied applied an an exponential exponential functi functionon to to define define the the nonlinear nonlinear verticalrate of adjustment vertical adjustment (b) of the ( Tedorib) of the River, Tedori Japan. River, Because Japan. the BecauseZmin of the each Zmin phase of each was notphase subtracted was not rate of vertical adjustment (b) of the Tedori River, Japan. Because the Zmin of each phase was not from the mean riverbed elevation prior to conducting nonlinear regression analysis, b tended to reach subtractedsubtracted from from the the mean mean riverbed riverbed elevation elevation prior prior to to conducting conducting nonlinear nonlinear regression regression analysis, analysis, b b zero with distance downstream and could not fully describe the trend of vertical adjustment. tendedtended to to reach reach zero zero with with distanc distance edownstream downstream and and could could not not fully fully describe describe the the trend trend of of vertical vertical adjustment. 4.3.adjustment. Lateral Variation

4.3.4.3. LateralTo Lateral investigate Variation Variation the transversal variation of riverbed elevation, four representative cross-sections located at the middle of each subreach are shown (Figure6). The horizontal line represents the To investigate the transversal variation of riverbed elevation, four representative cross-sections bankfullTo level.investigate The firstthe transversal notable point variation is that of there riverbed was elevation, only degradation four representative throughout cross the-sections entire located at the middle of each subreach are shown (Figure 6). The horizontal line represents the extentlocated of aat cross-section the middle of in each the period subreach 1950–1979, are shown while (Figure both 6 degradation). The horizontal and aggradation line represents took the bankfull level. The first notable point is that there was only degradation throughout the entire extent placebankfull simultaneously level. The first in dinotablefferent point parts is ofthat a there given was cross-section only degradation over the throughout period 1979–2007. the entire extent The of a cross-section in the period 1950–1979, while both degradation and aggradation took place synchronousof a cross-section degradation in the in period the former 1950– period1979, while could both be closely degradation related to and extensive aggradation SGM took that was place simultaneously in different parts of a given cross-section over the period 1979–2007. The synchronous conductedsimultaneously in the wholein different bed. parts One reason of a given for thecross latter-section is the over total the absence period 1979 of SGM–2007. from The 1991 synchronous and the degradation in the former period could be closely related to extensive SGM that was conducted in transversaldegradation nonuniformity in the former of period erosion could and depositionbe closely inducedrelated to by extensive the irregularitySGM that of thewas flow conducted velocity in thethe whole whole bed. bed. One One reason reason for for the the latter latter is is the the total total absence absence of of SGM SGM from from 199 1991 1and and the the transversal transversal nonuniformitynonuniformity of of erosion erosion and and deposition deposition induced induced by by the the irregularity irregularity of of the the flow flow velocity velocity field field in in

WaterWater2019 2019,,11 11,, 1852 x FOR PEER REVIEW 99 ofof 1717

which either scouring or settling of sediment particles can prevail in different parts of a given river field in which either scouring or settling of sediment particles can prevail in different parts of a given channel cross-section [14]. In reality, a change in structure of the flow velocity field along the cross- river channel cross-section [14]. In reality, a change in structure of the flow velocity field along the sectional direction was attributed to several factors, such as seasonal variation in flow discharge and cross-sectional direction was attributed to several factors, such as seasonal variation in flow discharge sediment supply, and the presence and destruction of turbulent eddy currents. This process led to and sediment supply, and the presence and destruction of turbulent eddy currents. This process led to the development of bars and islands in the river channel. the development of bars and islands in the river channel.

Figure 6. Superimposed cross-sectional profiles of the Tedori River at four cross-sections for 1950, Figure 6. Superimposed cross-sectional profiles of the Tedori River at four cross-sections for 1950, 1979, and 2007. (River kilometers mentioned here are referred from the river mouth. Elevation was 1979, and 2007. (River kilometers mentioned here are referred from the river mouth. Elevation was measured relative to Tokyo Peil (TP), the standard ground elevation in Japan based on the mean sea measured relative to Tokyo Peil (TP), the standard ground elevation in Japan based on the mean sea level of Tokyo Bay). level of Tokyo Bay). 4.4. Variations in Bankfull Channel Width 4.4. Variations in Bankfull Channel Width The bankfull channel widths averaged over SR were computed for 1950, 1960, 1970, 1978, 1991, 1997,The and bankfull 2002, as shownchannel in widths Figure 7averaged. The average over SR channel were widthcomputed of the for entire 1950, study 1960, reach1970, decreased1978, 1991, from1997, 330and m2002, in 1950as shown to 300 in mFigure in 2002. 7. The The average temporal channel variation width inof the channel entire width study indicatesreach decreased that a significantfrom 330 m decrease in 1950 in to channel 300 m widthin 2002. took The place temporal simultaneously variation in in channel the four width subreaches indicates from that 1950 a tosignificant 1970, coinciding decrease within channel the trend width of took considerable place simultaneously erosion in this in periodthe four along subreaches the LTR. from Similarly, 1950 to a1970, rapid coinciding decrease with in channel the trend width of considerable was still observed erosion in in this SR period 3 during along 1970–1978. the LTR. Similarly, Contrary a to rapid this phenomenon,decrease in channel channel width width was increased still observed by an averagein SR 3 during of 15 m 1970 in SR–1978. 1 and Contrary SR 2 as ato direct this phenomenon, consequence ofchannel the aggradation width increased trend from by an 1970 average to 1978. of Channel 15 m in width SR 1 in and the SR entire 2 as reach a direct slightly consequence decreased of from the 1978–1991aggradation and trend subsequently from 1970 remainedto 1978. Channel stable during width thein the last entire period, reach 1991–2002. slightly decreased from 1978– 1991 and subsequently remained stable during the last period, 1991–2002. 4.5. Variation in Sediment Characteristics 4.5. Variation in Sediment Characteristics Longitudinal variation in sediment diameter was examined next. A general downstream reduction in grainLongitudinal size accompanying variation a in general sediment decrease diameter in bed was slope examined was observed next. A (Figure general8). The downstream decline inreduction grain size in grain also reflects size accompanying the attrition ofa general gravel indecrease transit in along bed theslope rivers was [observed15]. During (Figure 1965–1975, 8). The sedimentdecline in meangrain diametersize also reflects (dm) increased the attrition in the of reach gravel of in 4.5–14 transit km, along ranging the rivers from [15]. 150 toDuring 400 mm 1965 in– the1975, river sediment channel mean and diameter 100 to 200 (d mmm) increased in the lateral in the area. reach From of 4.5 1979–14 km, to 1993, rangingdm for from the 150 river to channel400 mm decreasedin the river to channel 50–100 and mm, 100 while to 200 on mm the lateralin the lateral area, d area.m coarsened From 1979 in ato range 1993, fromdm for 200 the toriver 400 channel mm in thedecreased reach of to 3–14 50–100 km. mm, Similar while to on previous the lateral studies area, on dm SGM coarsened effects in [16 a, 17range], it wasfrom considered 200 to 400 mm that in SGM the reach of 3–14 km. Similar to previous studies on SGM effects [16,17], it was considered that SGM was responsible for the coarser dm value in the riverbed for two reasons: first, an amount of sand- and

Water 2019, 11, 1852 10 of 17 Water 2019, 11, x FOR PEER REVIEW 10 of 17 Water 2019, 11, x FOR PEER REVIEW 10 of 17 gravelwas responsible-sized grain for was the moved coarser outdm ofvalue the river, in the leaving riverbed the for bed two sediment reasons: layer ﬁrst, anwith amount larger ofgrains; sand- and and gravel-sized grain was moved out of the river, leaving the bed sediment layer with larger grains; and second,gravel-sized SGM graintriggered was erosion moved outupstream of the river,and downstream leaving the bedof the sediment excavation layer pit, with and larger finer grains; sediment and second, SGM triggered erosion upstream and downstream of the excavation pit, and finer sediment wassecond, thus SGMtransported triggered to the erosion downstream upstream area. and downstream of the excavation pit, and ﬁner sediment was thus transported to the downstream area. was thus transported to the downstream area.

Figure 7. Temporal trends of average channel width in four subreaches. FigureFigure 7.7. TemporalTemporal trendstrends ofof averageaverage channelchannel widthwidth inin fourfour subreaches.subreaches.

(a) (b) (a) (b) FigureFigure 8. 8. TemporalTemporal and and spatial spatial variations variations in sediment in sediment mean mean diameter diameter in (a) main in (a) course main courseand (b) andlateral (b ) Figure 8. Temporal and spatial variations in sediment mean diameter in (a) main course and (b) lateral arealateral. area. area. 4.6.4.6. Variation Variation in in Slope Slope 4.6. Variation in Slope AsAs a a consequence consequence of of the the direct direct effects effects of of human human activities activities or or changes changes in in other other controlling controlling factors, factors, As a consequence of the direct effects of human activities or changes in other controlling factors, thethe slope slope of of the the Tedori Tedori River River (S) (S) varied varied tempo temporallyrally and and spatially spatially at at different different rates rates during during the the period period the slope of the Tedori River (S) varied temporally and spatially at different rates during the period 19501950–2007–2007 ( (FigureFigure 9).9). Due Due to to unavailable unavailable topographical topographical data, data, a a gap gap appeared appeared in in the the period period 1980 1980–1990.–1990. 1950–2007 (Figure 9). Due to unavailable topographical data, a gap appeared in the period 1980–1990. ItIt was was found found that that variation variation in in channel channel slope slope in in SRs SRs 1 1 and and 4 4 was was more more significant significant than than that that in in othe otherr It was found that variation in channel slope in SRs 1 and 4 was more significant than that in other SRs.SRs. Commonly, Commonly, the the slope slope of of the the four four SRs SRs varied varied considerably considerably from from 1950 1950 to to 1991, 1991, and and subsequently subsequently SRs. Commonly, the slope of the four SRs varied considerably from 1950 to 1991, and subsequently remainedremained stable stable until until 2007. 2007. Over Over the the period period 1950 1950–1970,–1970, S Sof of SR SR 1 1 decreased decreased by by 30%, 30%, while while there there was was remained stable until 2007. Over the period 1950–1970, S of SR 1 decreased by 30%, while there was onlyonly a a slight slight variation variation in in S S in in SRs SRs 2 2 and and 3. 3. Possibly, Possibly, SGM SGM lowered lowered the the riverbed riverbed of of SR SR 3 3 by by the the same same only a slight variation in S in SRs 2 and 3. Possibly, SGM lowered the riverbed of SR 3 by the same magnitude,magnitude, and and consequently consequently S remained S remained at a at nearly a nearly constant constant level. level. According According to Dang to Danget al. [8], et al.from [8], magnitude, and consequently S remained at a nearly constant level. According to Dang et al. [8], from thefrom 1950s the to 1950s the to1960s, the 1960s,the amount the amount of SGM of decr SGMeased decreased by 50% by at 50% the atdownstream the downstream end of end SR of1 and SR 1 the 1950s to the 1960s, the amount of SGM decreased by 50% at the downstream end of SR 1 and increasedand increased by 10% by at 10% the at upstream the upstream end. Additionally, end. Additionally, SGM SGMupstream upstream decreased decreased the sediment the sediment load increased by 10% at the upstream end. Additionally, SGM upstream decreased the sediment load suppliedload supplied to SR to1. These SR 1. Theseactions actions decreased decreased S in SR S 1 in to SR establish 1 to establish a new aequilibrium new equilibrium condition. condition. Lane supplied to SR 1. These actions decreased S in SR 1 to establish a new equilibrium condition. Lane [18]Lane concluded [18] concluded that when that whena sediment a sediment load decreases load decreases from froman equil anibrium equilibrium condition condition in a stream, in a stream, the [18] concluded that when a sediment load decreases from an equilibrium condition in a stream, the equilibriumthe equilibrium state state will willbe lost; be lost; a new a new state state of equilibrium of equilibrium can can then then be be restored restored in in the the form form of of reduced reduced equilibrium state will be lost; a new state of equilibrium can then be restored in the form of reduced slope.slope. The The behavior behavior observed observed in in this this study study was was consistent consistent with with the the finding finding of of Lane Lane [18]. [18]. From From 1950 1950 to to slope. The behavior observed in this study was consistent with the finding of Lane [18]. From 1950 to 1972,1972, S S in in SR 4 increased byby 15%15% becausebecause there there were were SGM SGM activities activities in in the the downstream downstream portion portion but but not 1972, S in SR 4 increased by 15% because there were SGM activities in the downstream portion but not in the upstream. Subsequently, the opposite trend was observed in SR 4 as a result of laws not in the upstream. Subsequently, the opposite trend was observed in SR 4 as a result of laws

Water 2019, 11, 1852 11 of 17

Water 2019, 11, x FOR PEER REVIEW 11 of 17 Water 2019, 11, x FOR PEER REVIEW 11 of 17 in the upstream. Subsequently, the opposite trend was observed in SR 4 as a result of laws preventing preventing the overmining of sand and gravel in the mid-1970s. A change in the S, therefore, could thepreventing overmining the of overmining sand and gravel of sand in theand mid-1970s.gravel in the A mid change-1970s. in theA change S, therefore, in the couldS, therefore, be considered could be considered as a response of the Tedori River to anthropogenic activities, especially SGM. as abe response considered of theas a Tedori response River of the to anthropogenicTedori River to activities,anthropogenic especially activities, SGM. especially SGM.

Figure 9. Spatial and temporal variations in slopes. FigureFigure 9. 9.Spatial Spatial and and temporal temporal variationsvariations in slopes. 4.7. Variation in Total Stream Power and Specific Stream Power 4.7.4.7. Variation Variation in in Total Total Stream Stream Power Power and and Specific Specific Stream Stream Power Power Temporal variations in both TSP and SSP were investigated for four SRs of the LTR. There was TemporalTemporal variations variations in in both both TSP TSP and and SSP SSP were were investigatedinvestigated for four four SRs SRs of of the the LTR. LTR. There There was was similarity between variations in S and TSP during the period 1950–1979: TSP of SR 2 and SR 3 was similaritysimilarity between between variations variations in in S S and and TSP TSP during during thethe periodperiod 1950–1979:1950–1979: TSP TSP of of SR SR 2 2and and SR SR 3 3was was nearly unchanged, while TSP of SR 4 increased by 17% and TSP of SR 1 decreased by 15% during the nearlynearly unchanged, unchanged, while while TSP TSP of of SR SR 4 4 increased increased by by 17% 17% andand TSPTSP of SR 1 1 decreased decreased by by 15% 15% during during the the period 1950–1979 (Figure 10a). Variation in TSP of an SR was mainly attributed to an imbalance in periodperiod 1950–1979 1950–1979 (Figure (Figure 10 10a).a). Variation Variation in in TSP TSP ofof anan SRSR waswas mainly attributed attributed to to an an imbalance imbalance in in the amount of SGM at the upstream and downstream ends of the SR. In addition, TSP of four SRs thethe amount amount of of SGM SGM at at the the upstream upstream and and downstream downstream endsends ofof the SR. In In addition, addition, TSP TSP of of four four SRs SRs decreased significantly due to a reduction in mean annual flood discharge by 35% induced by TDO decreased significantly due to a reduction in mean annual flood discharge by 35% induced by TDO decreasedsince 1980. significantly TSP was nearly due tosteady a reduction from 1991 in to mean 2007 annual at four floodSRs as discharge a result of byno 35%SGM induced in this period. by TDO since 1980. TSP was nearly steady from 1991 to 2007 at four SRs as a result of no SGM in this period. sincePrior 1980. to 1991, TSP the was variation nearly steady trend in from SSP 1991was identical to 2007 at to four that SRsin the as S a and result TSP. of Since no SGM 1991, in except this period. for Prior to 1991, the variation trend in SSP was identical to that in the S and TSP. Since 1991, except for PriorSR to4, SSP 1991, at thethe variationother three trend SRs stood in SSP at was stable identical levels. The to that SSP in of the SR S4 andwas much TSP. Since higher 1991, andexcept fluctuated for SR SR 4, SSP at the other three SRs stood at stable levels. The SSP of SR 4 was much higher and fluctuated 4, SSPmore at than the that other of three the other SRs SRs. stood The at SSP stable in SR levels. 4 was The highest SSP ofin SR1970 4 wasbecause much of the higher smallest and bankfull fluctuated more than that of the other SRs. The SSP in SR 4 was highest in 1970 because of the smallest bankfull morewidth than together that of with the other high SRs.slope The (Figure SSP 10b). in SR The 4 was discrepancy highest in between 1970 because TSP of of two the consecutive smallest bankfull SRs width together with high slope (Figure 10b). The discrepancy between TSP of two consecutive SRs widthincreased together with with distance high downstream. slope (Figure In 10 contrast,b). The discrepancythe discrepancy between between TSP SSP of twoof two consecutive consecutive SRs increased with distance downstream. In contrast, the discrepancy between SSP of two consecutive increasedSRs decreased with distance with distance downstream. downstream. In contrast, the discrepancy between SSP of two consecutive SRs SRs decreased with distance downstream. decreased with distance downstream.

(a) (b) (a) (b) Figure 10. Spatial and temporal variations in (a) total stream power (TSP) and (b) specific stream FigureFigure 10. 10.Spatial Spatial and and temporal temporal variations variations in ( ain) total(a) total stream stream power power (TSP) (TSP) and (andb) speciﬁc (b) specific stream stream power power (SSP). (SSP).power (SSP).

Water 2019, 11, x FOR PEER REVIEW 12 of 17

4.8. Relationship between Vertical Variation of Riverbed Elevation and Related Bankfull Width The correlation between vertical adjustment of the average elevation of cross-sections and Water 2019changes, 11, 1852 of bankfull channel width in the LTR was examined for four SRs from 1950 to 2002 (Figure12 of 17 11). It was observed that vertical adjustment was not statistically correlated (Spearman’s rank correlation p >> 0.05) with the corresponding lateral adjustment at four SRs. In SR 1, some sections 4.8. Relationshipshowed channel between incision Vertical together Variation with of Riverbednarrowing, Elevation while others and Related showed Bankfull widening Width and channel deposition. A lower riverbed elevation made the flow more concentrated in the main course of the Thechannel, correlation leading between to a decreased vertical bankfull adjustment channel of width. the average In contrast, elevation an increased of cross-sections riverbed elevation and changes of bankfullmade the channel flow more width dispersed, in the LTRinducing was an examined increased forbankfull four SRschannel from width. 1950 Similarly, to 2002 channel (Figure 11). It wasincision observed and that narrowing vertical simultaneously adjustment was occurred not statistically during four correlatedphases at SR (Spearman’s 3. This coupled rank tendency correlation p >> 0.05of the) withLTR is the similar corresponding to that of some lateral Italian adjustment rivers, including at four the SRs.Piave, Inthe SR Po, 1, and some the Tevere sections [19]. showed In channelthis incision case, the together riverbank with protection narrowing, implemented while others earlier showed was considered widening and to be channel effective deposition. as a A lowercountermeasure riverbed elevation against made the instability the flow of more the concentrated riverbank induced in the by main river course degradation. of the channel, The leadingembankment to a decreased confined bankfull the river channel channel width., resulting In in contrast, no widening an increased due to bank riverbed failure in elevation the LTR in made SR 1 and SR 3. The early presence of bank protection along the LTR might have triggered the erosion the flow more dispersed, inducing an increased bankfull channel width. Similarly, channel incision in the riverbed. and narrowing simultaneously occurred during four phases at SR 3. This coupled tendency of the Besides the coupled tendency observed at SR 1 and SR 3, SR 2 and SR 4 still experienced channel LTR iswidening similar toand that incision of some at the Italian same time. rivers, This including phenomenon the Piave,can be explained the Po, and by theconsidering Tevere [that19]. the In this case, thesediment riverbank supply protection coming from implemented upstream reaches earlier was was decreased considered due to to be the e fflowective connectivity as a countermeasure with the againstmountain the instability reach. In of addition, the riverbank the severe induced incision by could river have degradation. induced a Thevery embankmentcoarse sublayer confined (Figure the river channel,9), and the resulting bed was armored. in no widening This mechanism due to led bank to erosion failure of in the the banks LTR rather in SR than 1 and further SR 3. incision The early presenceof the of channel. bank protection along the LTR might have triggered the erosion in the riverbed.

FigureFigure 11. Relationship 11. Relationship between between change change in in riverbed riverbed elevationelevation and and change change in inbankfull bankfull width. width. (Red (Red meansmeans phase phase I; purple I; purple means means phase phase II; II; green green means means phase phase III; blue III; means blue meansphase IV; phase black IV;means black phase means phaseV) V)..

Besides4.9. Role theof SSP coupled in Vertical tendency Adjustment observed at SR 1 and SR 3, SR 2 and SR 4 still experienced channel wideningSpatial and incision variation at in the SSP same and time.its association This phenomenon with vertical canadjustment be explained are shown by consideringin Figure 12. The that the sedimentresults supply indicate coming that SSP from was upstream scattered in reaches SR 3 and was SR decreased 4 but concentrated due to the in SR low 1 connectivityand SR 2. It seems with the mountainthat erosion reach. was In addition, featured in the SR severe 3 and incisionSR 4 when could SSP was have greater induced than a 231 very W/m coarse2. With sublayer SSP less (Figure than 9), 2 and the231 bed W/m was, both armored. deposition This and mechanism incision were led toobserved erosion at of SR the 1 and banks SR rather2. The thanSSP threshold further incision of 60 of W/m2 could be considered as the minimum energy to trigger the formation of erosion features. This the channel. value is higher than the threshold of transition from erosion to deposition between 30 and 35 W/m2 4.9. Role of SSP in Vertical Adjustment Spatial variation in SSP and its association with vertical adjustment are shown in Figure 12. The results indicate that SSP was scattered in SR 3 and SR 4 but concentrated in SR 1 and SR 2. It seems that erosion was featured in SR 3 and SR 4 when SSP was greater than 231 W/m2. With SSP less than 231 W/m2, both deposition and incision were observed at SR 1 and SR 2. The SSP threshold of 60 W/m2 could be considered as the minimum energy to trigger the formation of erosion features. This value is higher than the threshold of transition from erosion to deposition between 30 and 35 W/m2 WaterWater2019 2019, ,11 11,, 1852 x FOR PEER REVIEW 1313of of 1717 published previously [20]. Statistical correlation between SSP and vertical adjustment was examined published previously [20]. Statistical correlation between SSP and vertical adjustment was examined using Spearman’s correlation, which is a statistical measure of the strength of a monotonic using Spearman’s correlation, which is a statistical measure of the strength of a monotonic relationship relationship between paired data. A strongly negative correlation ( = −0.9, p-value < 0.1) was between paired data. A strongly negative correlation (ρ = 0.9, p-value < 0.1) was observed at SR 3. observed at SR 3. No statistically significant correlation between− SSP and vertical adjustment was No statistically significant correlation between SSP and vertical adjustment was identified at the other identified at the other three SRs (p >> 0.1). Therefore, statistical correlation between SSP and vertical three SRs (p >> 0.1). Therefore, statistical correlation between SSP and vertical adjustment varied at adjustment varied at different SRs. different SRs.

FigureFigure 12.12. RelationshipRelationship betweenbetween specificspecific streamstream powerpower (SSP)(SSP) andand vertical vertical adjustment. adjustment. (Red(Red meansmeans phasephase I;I; purplepurple meansmeans phase phase II; II; green green means means phase phase III; III; blue blue means means phase phase IV; IV; black black means means phase phase V). V). 4.10. Relationship between TSP and Sediment Diameter 4.10. Relationship between TSP and Sediment Diameter Variation in median sediment diameter (dm) in both the river channel and the flood plain was Variation in median sediment diameter (dm) in both the river channel and the flood plain was examined in relation to SGM, based on available annual surveys of sediment samples from 1963 to examined in relation to SGM, based on available annual surveys of sediment samples from 1963 to 1971. Figure 13 shows temporal variation in dm in relation to temporal variation in TSP. In SRs 1, 2, 1971. Figure 13 shows temporal variation in dm in relation to temporal variation in TSP. In SRs 1, 2, and 3 (Figure 13a–c), temporal variation in both dmc (median sediment diameter in the river channel) and 3 (Figure 13a–c), temporal variation in both dmc (median sediment diameter in the river channel) and dmp (median sediment diameter in the lateral area) is closely related to temporal variation in TSP and dmp (median sediment diameter in the lateral area) is closely related to temporal variation in TSP during 1963–1971. In SR 4 (Figure 13d), the variation in dmp had no relation to TSP from 1966 to 1969, during 1963–1971. In SR 4 (Figure 13d), the variation in dmp had no relation to TSP from 1966 to 1969, although a close relationship was also observed between dmc and TSP from 1963 to 1971. In particular, although a close relationship was also observed between dmc and TSP from 1963 to 1971. In particular, in 1966 and 1968, when the minimum values of TSP were measured, maximum values of dmp were in 1966 and 1968, when the minimum values of TSP were measured, maximum values of dmp were recorded (Figure 13d). The results of Spearman’s correlation analysis between TSP and dmc and dmp recorded (Figure 13d). The results of Spearman’s correlation analysis between TSP and dmc and dmp revealed no statistical correlation between them (|ρ| < 0.5, p >> 0.1). This unusual relationship could be revealed no statistical correlation between them (|| < 0.5, p >> 0.1). This unusual relationship could related to the effect of intensive SGM in the lateral area of SR 4, as gravel and sand extraction might be related to the effect of intensive SGM in the lateral area of SR 4, as gravel and sand extraction might have made the bed surface more armored. have made the bed surface more armored.

4.11. Relationship between River Pattern and other River Characteristics Finally, variation in the river configuration of the LTR was observed based on sequential aerial photographs at a scale of 1:15,000 for 1955, 1968, 1984, 1995, and 2000 (only photographs for 1955 and 1968 are shown in Figure 14). Commonly, a braided pattern was observed. In general, the braided pattern of a river is formed as a result of local sorting and deposition of coarse material that cannot be locally transported. The coarse material grows into a central bar and subsequently an island [21]. The braided pattern of the LTR was studied based on slope, bankfull discharge, and mean sediment diameter. According to the studies of Leopold and Wolman [22] and Henderson [23], the river pattern becomes braided if the river slope (S) is greater than the critical slope (S*):

Water 2019, 11, x FOR PEER REVIEW 14 of 17

S > S* (4) Here, based on the available data of slopes, mean annual and bankfull discharge, and mean sediment diameter, the patterns of the Tedori River were estimated according to Equation (4) for several periods (Table 1). From Table 1, it was deduced that, despite significant degradation due to WaterSGM,2019 the, 11LTR, 1852 had a braided pattern in all examined periods. 14 of 17

(a) (b)

Figure 13. TemporalTemporal variation in total stream power and sediment mean diameter (dm)) of of lower lower Tedori Tedori River (LTR).

4.11. Relationship between River Pattern and Other River Characteristics Finally, variation in the river conﬁguration of the LTR was observed based on sequential aerial photographs at a scale of 1:15,000 for 1955, 1968, 1984, 1995, and 2000 (only photographs for 1955 and 1968 are shown in Figure 14). Commonly, a braided pattern was observed. In general, the braided pattern of a river is formed as a result of local sorting and deposition of coarse material that cannot be locally transported. The coarse material grows into a central bar and subsequently an island [21]. The braided pattern of the LTR was studied based on slope, bankfull discharge, and mean sediment diameter. According to the studies of Leopold and Wolman [22] and Henderson [23], the river pattern becomes braided if the river slope (S) is greater than the critical slope (S*):

S (>a)S* (4)

Here, based on the available data of slopes, mean annual and bankfull discharge, and mean sediment diameter, the patterns of the Tedori River were estimated according to Equation (4) for several periods (Table1). From Table1, it was deduced that, despite signiﬁcant degradation due to SGM, the

LTR had a braided pattern in all examined periods. Water 2019, 11, x FOR PEER REVIEW 14 of 17

S > S* (4) Here, based on the available data of slopes, mean annual and bankfull discharge, and mean sediment diameter, the patterns of the Tedori River were estimated according to Equation (4) for several periods (Table 1). From Table 1, it was deduced that, despite significant degradation due to SGM, the LTR had a braided pattern in all examined periods.

(a) (b)

Water 2019Figure, 11 ,13. 1852 Temporal variation in total stream power and sediment mean diameter (dm) of lower Tedori15 of 17 River (LTR).

Water 2019, 11, x FOR PEER REVIEW 15 of 17 (a)

(b)

FigureFigure 14.14. Comparison ofof aerial photographsphotographs ofof LTRLTR overover twotwo dates.dates.

TableTable 1.1. PatternsPatterns ofof thethe TedoriTedori RiverRiver forfor severalseveral periods.periods.

PatternsPatterns of Tedori of Tedori River River No.No. PeriodPeriod 0.44 −0.44 1.15 0.461.15 −0.46 S* = 0.013S* = (Q 0.013b)− (Q[22b)] [22]S* = 0.002S* d=50 0.002Q db50− Q[23b ] [23] 11 1950–19611950–1961 BraidedBraided N/A N/A 22 1962–19711962–1971 BraidedBraided BraidedBraided 33 1972–19791972–1979 BraidedBraided N/A N/A 1991, 1993, 1991, 1993, 1997,1998, 4 1997,1998, Braided N/A 4 1999, 2002, Braided N/A 2003,1999, 2007 2002,

Note: Dates were divided based2003, on 2007 the availability of data including slopes and median grain size of sediment. Qb, 3 3 bankfullNote: Dates discharge were (1144 divided m /s forbased post-dam on the period availability (1980–2007) of data and including 1755 m /s for slopes pre-dam and period median (1950–1979)); grain sized 50of, median grain size of sediment in the river channel; N/A, not available. sediment. Qb, bankfull discharge (1144 m3/s for post-dam period (1980–2007) and 1755 m3/s for pre-

dam period (1950–1979)); d50, median grain size of sediment in the river channel; N/A, not available. Overall, a series of ﬁndings obtained in this study through comprehensive inter-correlated analysis are of signiﬁcance for a better understanding of how river channels respond to human impacts such as Overall, a series of findings obtained in this study through comprehensive inter-correlated SGM and dam operation. analysis are of significance for a better understanding of how river channels respond to human 5.impacts Conclusions such as SGM and dam operation.

5. ConclusionsA better understanding of changes in river morphology is fundamental to enable the sustainable management of river basins. The present study conducted a comprehensive analysis of a 58-year A better understanding of changes in river morphology is fundamental to enable the sustainable topographic survey of the LTR in Japan and related data on human impact to clarify the long-term management of river basins. The present study conducted a comprehensive analysis of a 58-year variation in morphological characteristics. The empirical model successfully described vertical topographic survey of the LTR in Japan and related data on human impact to clarify the long-term adjustment of the LTR in ﬁve phases. The nonlinear rate extracted from the empirical model fully variation in morphological characteristics. The empirical model successfully described vertical captured the characteristics of vertical adjustment of the LTR. The principal features observed in this adjustment of the LTR in five phases. The nonlinear rate extracted from the empirical model fully study are summarized as follows: Degradation occurred at SR 3 and SR 4 over all phases. Aggradation captured the characteristics of vertical adjustment of the LTR. The principal features observed in this and degradation phases appeared alternately at SR 2. Aggradation phases following degradation study are summarized as follows: Degradation occurred at SR 3 and SR 4 over all phases. phases were observed at SR 1. These are the typical responses of gravel-bed rivers in Japan where SGM Aggradation and degradation phases appeared alternately at SR 2. Aggradation phases following has been prohibited. It is also shown that variation in bankfull width was accompanied by variation degradation phases were observed at SR 1. These are the typical responses of gravel-bed rivers in in riverbed elevation; speciﬁcally, decreased bankfull width was generally coupled with riverbed Japan where SGM has been prohibited. It is also shown that variation in bankfull width was accompanied by variation in riverbed elevation; specifically, decreased bankfull width was generally coupled with riverbed degradation, and increased bankfull width was coupled with riverbed aggradation. The river pattern is deduced to have stayed braided during the period 1950–2000 as a result of mutual interaction among high slopes, mean sediment diameter, and bankfull discharge. The findings obtained in the present study will provide basic scientific/engineering information for the effective management of the river. The obtained results are also useful as they can be used to understand and compare the behavior of other rivers of a similar type over the world.

Water 2019, 11, 1852 16 of 17 degradation, and increased bankfull width was coupled with riverbed aggradation. The river pattern is deduced to have stayed braided during the period 1950–2000 as a result of mutual interaction among high slopes, mean sediment diameter, and bankfull discharge. The findings obtained in the present study will provide basic scientific/engineering information for the effective management of the river. The obtained results are also useful as they can be used to understand and compare the behavior of other rivers of a similar type over the world.

Author Contributions: The authors designed the framework for the analysis together: D.M.H. analyzed the data and wrote the manuscript; S.U. collected the data and provided suggestions and improvements to the manuscript; M.Y. guided the study and provided the suggestions and improvements to the manuscript. All authors read and approved the manuscript. Funding: The study was partially supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. Acknowledgments: The authors are grateful to the Hokuriku Regional Development Bureau, Ministry of Land, Infrastructure, Transport, and Tourism of Japan, for providing the field data. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Simon, A. A model of channel response in disturbed alluvial channels. Earth Surf. Process. Landf. 1989, 14, 11–26. [CrossRef] 2. Rinaldi, M.; Simon, A. Bed-level adjustments in the Arno River, central Italy. Geomorphology 1998, 22, 57–71. [CrossRef] 3. Oguchi, T.; Saito, K.; Kadomura, H.; Grossman, M. Fluvial geomorphology and paleohydrology in Japan. Geomorphology 2001, 39, 3–19. [CrossRef] 4. Yoshimura, C.; Omura, T.; Furumai, H.; Tockner, K. Present state of rivers and streams in Japan. River Res. Appl. 2005, 21, 93–112. [CrossRef] 5. Huang, G. Time lag between reduction of sediment supply and coastal erosion. Int. J. Sediment Res. 2011, 26, 27–35. [CrossRef] 6. Sato, S.; Kajimura, T.; Abe, M.; Isobe, M. Sand Movement and Long-Term Beach Evolution in a Fluvial System Composed of the Samegawa River and the Nakoso Coast. Coast. Eng. J. 2004, 46, 219–241. [CrossRef] 7. Dang, M.H.; Umeda, S.; Yuhi, M. Long-term riverbed response of lower Tedori River, Japan, to sediment extraction and dam construction. Environ. Earth Sci. 2014, 72, 2971–2983. [CrossRef] 8. Dang, M.H.; Yuhi, M.; Umeda, S. Human Impact on Morphology and Sediment Budget in the Tedori River, Japan; Advances in River Sediment Research; CRC Press: Boca Raton, FL, USA, 2013; pp. 289–297. 9. Teramoto, A.; Nakamura, S.; Tsujimoto, T. Change of bar formation and bank erosion in the Tedori River. Adv. River Eng. JSCE 2003, 9, 329–334. 10. Yamamoto, Y.; Kawai, K.; Takan, T. Sedimentation control of coastal zone with large-scale sandy beach. Proc. Civ. Eng. Ocean JSCE 2008, 24, 1339–1344. (In Japanese) [CrossRef] 11. Leopold, L.B. A View of the River; Harvard University Press: Cambridge, MA, USA, 1994. 12. Jang, C.-L.; Shimizu, Y. Numerical Simulation of Relatively Wide, Shallow Channels with Erodible Banks. J. Hydraul. Eng. 2005, 131, 565–575. [CrossRef] 13. Bagnold, R.A. Bed load transport by natural rivers. Water Resour. Res. 1977, 13, 303–312. [CrossRef] 14. Alekseevskiy, N.; Berkovich, K.; Chalov, R. Erosion, sediment transportation and accumulation in rivers. Int. J. Sediment Res. 2008, 23, 93–105. [CrossRef] 15. Collins, B.D.; Dunne, T. Gravel transport, gravel harvesting, and channel-bed degradation in rivers draining the southern olympic mountains, Washington, U.S.A. Environ. Earth Sci. 1989, 13, 213–224. [CrossRef] 16. Kondolf, G.M. Geomorphic and environmental effects of instream gravel mining. Landsc. Urban Plan. 1994, 28, 225–243. [CrossRef] 17. Rinaldi, M.; Wy˙zga,B.; Surian, N. Sediment mining in alluvial channels: Physical effects and management perspectives. River Res. Appl. 2005, 21, 805–828. [CrossRef] 18. Lane, E.W. The importance of fluvial morphology in hydraulic engineering. Am. Soc. Civ. Eng. Proc. 1955, 81, 1–17. Water 2019, 11, 1852 17 of 17

19. Surian, N.; Rinaldi, M. Morphological response to river engineering and management in alluvial channels in Italy. Geomorphology 2003, 50, 307–326. [CrossRef] 20. Bizzi, S.; Lerner, D.N. The Use of Stream Power as an Indicator of Channel Sensitivity to Erosion and Deposition Processes. River Res. Appl. 2015, 31, 16–27. [CrossRef] 21. Akhtar, M.; Sharma, N.; Ojha, C. Braiding process and bank erosion in the Brahmaputra River. Int. J. Sediment Res. 2011, 26, 431–444. [CrossRef] 22. Leopold, L.B.; Wolman, M.G. River Channel Pattenrs: Rraided, Meandering and Straight; Physiographic and Hydraulic Studies of Rivers, Geological Survey Professional Paper 282-B; United States Government Pringing Oﬃce: Washington, DC, USA, 1957; p. 85. 23. Henderson, F.M. Stability of Alluvial Channels. J. Hydraul. Div. 1963, 87, 109–138.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).