Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

1 Modeling hydrodynamic changes induced by Run-of-River hydropower plants

2 along Prahova River (Romania)


3
4 Journal of Energy Engineering, ASCE, 2018, 144(2): 04017078,
5 DOI: 10.1061/(ASCE)EY.1943-7897.0000515
6
7 Daniela Elena Gogoaşe Nistoran1; Cristina Sorana Ionescu2; Livioara Braşoveanu3;
8 Iuliana Armaş 4; Ioana Opriş5 and Sorina Costinaş6
9
1
10 Associate Professor, PhD, M. Faculty of Power Engineering, Department of
11 Hydraulics, Hydraulic machines and Environmental engineering, University
12 Politehnica of Bucharest, 313 Spl. Independentei, Bucharest, 060042, Romania,
13 (corresponding author); ASCE Member between 2004-2013, Email:
14 dnistoran@gmail.com, daniela.nistoran@upb.ro
2
15 Associate Professor, PhD, Faculty of Power Engineering, Department of Hydraulics,
16 Hydraulic machines and Environmental engineering, University Politehnica of
17 Bucharest, 313 Spl. Independentei, Bucharest, 060042, Romania; Email:
18 cristina.sorana.ionescu@gmail.ro, cristina.ionescu@upb.ro
3
19 PhD student, Faculty of Geography, University of Bucharest, no. 1, N. Balcescu bd.,
20 Bucharest, 010041, Romania; Email: brasoveanulivioara@yahoo.com
4
21 Professor, PhD, Faculty of Geography, University of Bucharest, no. 1, N. Balcescu
22 bd., Bucharest, 010041, Romania; Email: iulia_armas@geounibuc.ro
5
23 Associate Professor, PhD, Faculty of Power Engineering, Department of Energy
24 Generation and Use, University Politehnica of Bucharest, Bucharest, 060042,
25 Romania; Email: ioana.opris@upb.ro
6
26 Professor, PhD, Faculty of Power Engineering, Department of Energy Generation
27 and Use, University Politehnica of Bucharest, Bucharest, 060042, Romania; Email:
28 Email: sorina.costinas@upb.ro
29
30 ABSTRACT
31 The purpose of the study is to reveal changes of the hydrodynamics and sediment
32 transport induced into the river channel by the operation of small hydropower plants
33 (SHP). A 36 km reach of the upper Prahova River (Romania) has been chosen, along
34 which 10 run-of-river developments are built. HEC-RAS hydraulic modeling
35 software was used to compute hydrodynamic parameters over the selected reach. Two
36 discharge scenarios were considered into the channel: (i) mean annual flow (for the
37 natural regime, without SHP) and (ii) the required minimum discharge (“worst case
38 scenario” of SHP operation for the environment). Results of the second scenario show
39 that ecologically relevant hydraulic parameters in the channel are approaching the
40 thresholds of endangering fish habitat. On average, depth and shear stress drop by
41 50%, mean velocity by 40% and flow area by 75%. Comparison between computed
42 nondimensional shear stress and its corresponding critical value indicate that
43 minimum discharge can only transport finer particles than average from the bed
44 surface, thus leading to potential filling of fish spawning areas with sand.
45
46 Keywords: environmental impact, fish habitat, HEC-RAS, hydraulic
47 modeling, run-of-river hydropower, sediment transport

1
48 INTRODUCTION
49
50 On average, hydropower represents 30% of the total power produced in Romania and
51 is considered to be one of the most reliable and cost-effective source of energy
52 (Bruno and Fried 2008, Stematiu 2008). The output of this renewable energy depends
53 on the water discharge, head, efficiencies of the hydraulic circuit, turbine, generator
54 and operation time.
55 As most of the large hydropower potential of the European and Romanian big rivers
56 has already been harnessed into large impoundment development schemes and the
57 demand for energy is growing, a significant resurgence of interest has appeared
58 during the last decades for small hydropower plants (SHPs) (Anderson et al. 2015).
59 Such developments are also largely seen as having lower environmental impact than
60 large-scale schemes (Punys et al. 2015, Kaunda et al. 2012). European Union (EU)
61 and national legislation pertaining to energy production from renewable sources has
62 been issued (European Parliament and Council of The European Union 2009,
63 Romanian Government 2007 a, Romanian Parliament 2008, Romanian Ministry of
64 Environment and Forests 2012) to support requirements for design, construction and
65 operation of SHPs. Introduction of “green” certificate incentives for the suppliers of
66 renewable energy has encouraged the development of SHPs in Romania, even though
67 environmental requirements are strict.
68 The run-of-river (RoR) type of SHP development schemes are tailored to each site.
69 Typically, they may be high-head schemes (harvesting low discharge values), or low
70 head schemes (harvesting large discharge values), generally involving a small weir to
71 divert the water to a canal and/or penstock pipes. RoR schemes are considered to
72 produce green energy (Demirabas 2007, Fraenkel et al. 1991) since they do not
73 produce waste and CO2 and usually have ecological intakes with small diversion
74 weirs and fish pass-facilities that do not affect river continuity. In addition, the layout
75 of the penstock pipes, coupled with abstraction of water from the regulated river,
76 have to comply with European (European Parliament and Council of The European
77 Union 2000) and national legislation (Romanian Parliament 1996, Romanian
78 Government 2007 b) to minimize their impact on the river ecological status. RoR
79 SHPs generally operate at the base of the load curve, as compared with the large
80 hydropower plants with big reservoirs, operating at the peak of the load curve (with
81 frequent switch-ons and shut-offs).
82 However, operation of the RoR plants without storage facilities is restricted by the
83 natural river discharge and the requirement to protect the downstream ecological life
84 and plant equipment. At very low flows, turbines have either to operate at partial load
85 or to stop, to allow the flowing of the minimum discharge (sometimes called “hands
86 off” flow), Qmin into the channel. This minimum discharge adds up the environmental
87 flow rate (namely Eflow, Qe, for preserving aquatic life) and the discharge required to
88 supply the downstream consumers water demand (European Commission 2015,
89 Romanian Parliament 1996) along the potentially flow depleted reach (PFDR),
90 between intake and channel restitution. At high flows, particularly during floods, the
91 SHPs must be carefully monitored so that the intake trash racks should not be clogged
92 with debris that could affect the turbine blades. SHP schemes have the greatest impact

2
93 on channel flow for mid-range and low flows (Punys 2015, Environmental Agency
94 2009).
95 Very little information exists regarding the influence of SHP development schemes
96 on the hydrodynamic and sediment transport regime over the PFDRs. (Anderson et al.
97 2006, Csiki and Rhoads 2010). Particularily, studies exploring differences in the
98 ecological status, vegetation and fish habitat relative to the natural environment of the
99 river are requested (Anderson et al., 2015, Benejam et al. 2016). According to
100 Benejam (2016), an ecological index to assess fish quality of life may be computed
101 based on three scores: velocity, depth and substrate.
102 Within this context, the objective of present paper is to analyse the changes of the
103 hydrodynamic parameters into a stream channel, that are associated with the diversion
104 of water through the penstock of RoR hydropower plants, relative to the average,
105 natural flow regime. Furthermore, to analyse the possible overall physical alteration
106 of aquatic environment, the paper investigates the impact of these changes on
107 sediment transport. A 1D hydraulic model is set up to compare the values of velocity,
108 stream depth, cross-section area and bed shear stress, for two steady flow scenarios:
109 mean annual flow rate, Qmean, and minimum allowable flow rate, Qmin. The Prahova
110 River from Romania was selected as an example, because its Carpathian and
111 Subcarpathian reaches were subjected to intense SHP development during the last
112 decades.
113
114 STUDY AREA
115 Prahova River is a first order tributary of Ialomita River and a second order tributary
116 of Danube River, having its source in the Romanian Carpathians (Bucegi Mountains).
117 Prahova River’s length is 193 km; it has a catchment area of 3750 km2 (Fig. 1) and an
118 average bed slope of about 5.4 m/km.
119

120
121 Fig. 1 Prahova River watershed and Valley; Small Hydropower Plants (SHP) and Gauging
122 Stations (GS) along upper reach
123
124 The 56km upper reach of this river (Fig. 1) is situated between the source and the
125 downstream confluence with main left tributary, Doftana River (taken as a reference

3
126 for chainage throughout the paper). Three gauging stations (GS) are at Campina (km
127 2), Poiana Ţapului (km 40) and Buşteni (km 45). Mean annual discharge values range
128 between 0.5 m3/s (at source) and 8.1 m3/s at the downstream end, while the average
129 bed slope is of 12m/km.
130 The Subcarpathian reach of upper Prahova River and its valley have been severe
131 anthropically modified, during the last two centuries by extensive gravel mining,
132 construction of multiple roads, bridges and railway, channel regulation and bank
133 protection works, building of small check dams and two larger dams across the river,
134 etc., thus affecting the river hydraulics and morphology As a consequence, the
135 channel along this reach (with very few tributaries) became sediment depleted,
136 incised and abandoned its secondary channels, transforming into a sinuous almost
137 single-thread planform pattern (Armaş et al. 2012, Toroimac et al 2010, Wohl 2006,
138 Latapie et al. 2014).
139 At the end of the 19th century the largest hydropower plant in the country was built at
140 Sinaia 0 (km 35), to which, one century later, in chronological, order were added
141 Sinaia 2 (km 32), Sinaia 3 (km 31), and Sinaia 1 (km 33) (Fig. 1 and Fig. 2), thus
142 forming a group of four RoR units.
143

144
145 Fig. 2. Layout scheme of the small hydropower plants along upper Prahova River

146 After 2010 new river regulation works started, together with the construction of two
147 new groups of small hydropower developments, each with three RoR units. At the
148 same time river engineering works in the channels were developed to protect the
149 collapsing banks (by gabions) and incising bed (by small elevation check dams or
150 stepped weirs with baffle blocks or stilling basins with sills). Due to natural river
151 slope and to these constructions, the flow regime in the stream channel has transitions
152 from subcritical to supercritical regimes.
153
154 DATA AND METHOD

4
155 Small Hydropower Plants
156 Along the downstream reach of upper Prahova River, starting at 36 km from the
157 confluence with Doftana tributary, three groups of 10 in-series RoR SHPs were built
158 (by two different developers) (Fig. 1, Fig. 2, Fig. 3 and Table 1). Two dams with
159 small reservoirs exist upstream Sinaia 0 SHP, and upstream Sinaia 2 SHP, (Fig. 2).
160 Other weirs with ecological intake structures (equipped with desilting basins, forebay
161 tanks, trash racks and fish-pass structures) are built at Sinaia (km 31), Posada (km 25)
162 and at Breaza-Belia (km 17.3) (Fig. 2).
163

164
165 Fig. 3 Longitudinal profile of study reach along Prahova with the 10 RoR SHPs
166
167 Table 1 Small Hydropower Plant Data
SHP name Commi Type of Turbine Gross Q mean Q min Q design P design
ssioning turbine units Head, (m3/s) (m3/s) (m3/s) (MW)
year H (m)
Sinaia 0 1898 Francis 4 20.1 6.6 0.90 5.2 1.0
Sinaia 1 1993 Francis 3 17.5 6.6 0.90 5.2 1.0
Sinaia 2 1988 Francis 3 16.0 6.9 1.00 6.0 1.0
Sinaia 3 1989 Francis 3 17.0 7.1 1.00 6.0 1.2
Group 1 4.2
Posada 2014 crossflow 2 93.3 7.3 1.10 7.0 5.2
Comarnic 2014 crossflow 2 62.0 7.4 1.10 7.0 3.5
Breaza 2014 crossflow 2 42.0 7.6 1.10 7.0 2.4
Group 2 11.1
Cornu 2011 crossflow 2 49.4 7.6 1.25 8.0 3.0
Lunca
2011 crossflow 2 47.6 7.8 1.25 8.0 2.9
Cornului
Campina 2012 crossflow 2 48.6 8.1 1.25 8.0 2.9
Group 3 8.8
168

5
169 The newest, downstream 6 SHPs are equipped with cross-flow turbines that can
170 operate efficiently even at partial load (design flow). To extend the operation time of
171 hydropower plants during low flow periods, more than one turbine unit was installed
172 per each SHP. Minimum abstracted discharge for the three downstream SHPs is 20%
173 of Qdesign 1 turbine.
174
175 Hydrological analysis
176 Mean daily discharge values registered during different periods of time at the three
177 gauging stations were analysed in order to characterise the hydrological regime of
178 upper Prahova River. Mean (multi) annual flow, Qmean, was computed and resulted to
179 be 8.1 m3/s at Campina (between 1964-2011), 4.92 m3/s at Poiana Ţapului GS
180 (between 1966-1992) and 3 m3/s at Busteni GS (between 1966-2011).
181

182
183 Fig. 4 Flow duration curve for Campina SHP (with operation of two individiual turbine units)
184
185 As an example, a flow duration curve (FDC) was computed at Campina GS from
186 mean daily discharge values during 48 years (Fig. 4). Qmean was found to correspond
187 to a probability of exceedance of about 36% (Qmean = Q36, which means Qmean is
188 statistically exceeded 130 days a year). Usually, discharges between Q10 and Qmean are
189 considered high flows (with flows between Q10 and Q1 considered extremely high
190 flows, for which normally SHP intake valves close), whereas discharges between Q70
191 and Q90 are considered low flows (with flows between Q90 and Q100 considered
192 extremely low flows) (Otuagoma et al. 2016). SHPs can practically be used to operate
193 from partial to full load for discharges between Q10 and Q70 (mid-range or moderate
194 flows, including freshes). The ratio Q50/Q90 shows a reduced variability of low flow
195 discharges for the considered river reach. The two horizontal fringes in Fig. 4
196 correspond to flow ranges for which the power plant may operate from partial to full
197 load, with one and both turbines.
198 SHP licence to abstract water from a river specifies that Qmin (reserved flow rate)
199 should remain in the stream. A wide range of methodologies have been developed to

6
200 assess the ecological flow (Gopal 2013, Huckstorf et. al. 2008, Arthington and
201 Zalucki 1998), such as hydrological, hydraulic rating, habitat simulation, holistic etc.
202 So far no universally valid solution exists (European Commission 2015, ESHA
203 2005). For the present paper the ecological flow rate, Qe, was computed as the
204 smallest of the two values: (i) 10% of the Qmean and (ii) the Q95 value that corresponds
205 to 95% probability of exceedance on the minimum monthly flow duration curve
206 (European Parliament and Council of The European Union 2000). For example, in the
207 case of Campina GS the smallest discharge of the two values: 0.81 m3/s and 1.1 m3/s
208 flows through the fish-pass structure and over the spillway (Fig. 5). Since no data was
209 available for the necessary discharge to satisfy consumers along the entire Sinaia-
210 Campina PFDR, this was estimated to be 50% of the Qe. Therefore, minimum Eflow
211 value from Table 1 was computed with
212 Qmin  Qe  50%Qe.  10%  Qmean  5%  Qmean  15%  Qmean (1)
213 for all SHPs. Since three SHPs operate in-series along the same penstock, the same
214 Qmin value, obtained for the downstream SHP unit within a group of developments
215 was considered for the entire individual PFDR, up to the common upstream intake
216 (Table 1).
217

218
219 Fig. 5 Monthly (minimum) flow duration curve (Qm.m.) at Campina GS; Qe = Q95
220
221 Sediments and Morphology
222
223 Wolman pebble count method (Leopold and Wolman 1957, Leopold et al. 1964) was
224 applied in 2012 to sample sediment particles from the surface layer (exposed on top
225 of the streambed), along the main channel of the subcarpathian PFDR. Results for the
226 five survey sub-reaches are presented in the sediment grading curve from Fig. 6,
227 showing a gravel-bed stream, with a a sediment mixture of sand, cobbles and
228 particularily gravel (Parker, 2008, Radoane et al. 2008). Similar values for the d50 of
229 about 28 mm (corresponding to very coarse gravel) in the surface layer were
230 registered for all of the 5 sampling subreaches.
231

7
232
233 Fig. 6 Sediment grading curves along study reach: pebble count from the bed surface and
234 sieving from the bed subsurface (in quarries)
235
236 No sediment measurements from the bed substrate were available for the present
237 paper. However, Pascu (1999) obtained two sediment grading curves (Fig. 6) by
238 sieving samples taken from two sediment quarries situated in the study area at Podu
239 Vadului (km 8.5) and Nistoresti (km 15). As it may be seen in Fig. 6, the mean
240 diameter of these substrate samples, of about 4 mm, is smaller than the corresponding
241 one of the surface layer.
242 Sediment gradation is very important for fish spawning, nursing and sheltering of
243 many species. Fish seem to prefer areas with coarser sediment stratum (gravel), in
244 which some species even dig holes with their tails for spawning. A streambed high in
245 silt load can result in increased anoxia by burial, which in turn leads to higher egg
246 mortality (Kondolf and Wolman 1993). Other important variables for sediment
247 transport are vertical sorting, packing and pore space of the substrate. However, for
248 the present paper, no attempts were made to analyse these factors.
249
250 Fish habitat requirements
251

8
252 Review of current literature on physical (hydromorphological) and ecological impact
253 of hydropower on rivers found most of papers deal only or mostly with fish (Robson
254 et al. 2011), neglecting other riverine aquatic biota. Even though the purpose of the
255 paper is only to reveal the changes of river channel hydrodynamics induced by
256 operation of RoR SHP, the authors believe engineers should also know which are the
257 ecologically relevant parameters and their threshold for a specific river reach.
258 According to fish communities (Angermeier and Davideanu, G., 2004) fish species
259 encountered along upper Prahova River and its tributaries belong to the Cyprinidae
260 and Salmonidae families. The most important of them are the nase (Chondrostoma
261 Nasus), barbel (Barbus Barbus), grayling (Thymallus Thymallus, protected species),
262 trout (Salmo Trutta Fario), and chub (Leuciscus cephalus).
263 All of these species have different physical factors and require certain tolerable values
264 of hydrodynamic parameters, temperature and stratum cover structure (depth and
265 coarseness), in order to maintain good habitat conditions for life, migration, spawning
266 and nursing (Ingram et. al. 1999).
267 There is a large amount of literature sorrounding fish habitat physical requirements,
268 depending on their dimension, behaviour or needs (Robson et al. 2011, Lamouroux
269 and Cattanéo 2006). Biological studies dealing with the aforementioned species
270 indicate the necessary values of physical parameters as being: depths from 20 cm to
271 80 cm (with even 45 cm-1.8 m for grayling), stream velocities ranging between 0.25-
272 1.2 m/s (with lower 0.25-0.6 m/s for grayling and higher 0.6-1.3 m/s for nase), at least
273 2-3 cm in thickness bed mixture consisting in boulders (>256 mm), cobbles (64-256
274 mm), but mostly gravels (2-64 mm), preferably with plant roots vegetation, and
275 temperatures between 4-19 C (with a larger span for trout, 1-22 C and a narrow
276 span for barbel, 10-19 C) (Vehanen et al. 2003, Melcher and Schmutz 2010,
277 Arlinghaus and Wolter 2003). From these ranges, minimum necessary values for fish
278 habitat along the studied PFDR should be: 0.2 m for depth, 0.25 m/s for velocity and
279 2 cm for substrate depth with a minimum d50 of 8-16 mm, coresponding to medium
280 gravel. However, further biological and ecological studies are necessary to investigate
281 in depth these values and aspects.
282
283 Hydraulic model
284
285 LiDAR topographic data of the Prahova River Valley and floodplain were coupled (in
286 ARCGIS) with channel bathymetric cross-section surveys to obtain an extended DTM
287 (as a TIN) of the study river strech (Fig. 1). HEC-GeoRAS (HEC 2011) utility was
288 used to extract from this DTM a schematic network with 140 cross-sections. The
289 geometry of a 1D hydraulic model was then imported into HEC-RAS software (HEC
290 2010, Nistoran et al. 2007, Gogoaşe Nistoran D.E. et al. 2017) where it was further
291 refined through interpolation at a maximum distance of 100 m, obtaining 636 cross-
292 sections (Fig. 7).
293

9
294
295 Fig. 7 3D view of the upper Prahova River channel (HEC-RAS hydraulic model)
296
297 Simulations were performed under the mixed (subcritical and supercritical) steady
298 flow regime, to take into account possible regime transitions. Inflow tributaries were
299 considered as local flow “injections” at confluences in the model, leading to step
300 variation of discharge with distance. Also local discharge abstractions/”injections”
301 were considered at intakes/tailraces along the study reach to account for the operation
302 of SHP. Flow rate values at the upstream end and at locations along stream where
303 variations occur were considered as boundary conditions, together with channel slope
304 values at the ends of the river reach.
305 Model calibration was performed without abstraction of flow for SHP operation in
306 terms of roughness coefficients on observed flow rating curves (Fig. 8) at the two
307 GSs of Câmpina and Buşteni for low discharge values: minimum ecological flow
308 (Qmin), mean annual discharge (Qmean) 5-year flood (which was considered equal to
309 the bankfull discharge), and the 10-year flood (with 10% probability of exceedance).
310 Manning roughness coefficients were nc = 0.031 for the main channel and nf = 0.062
311 for the floodplain.
312

10
313
314 Fig. 8 Comparison between observed (points) and computed (lines) rating curves at the
315 Campina GS and Busteni GS
316
317 Also, water surface elevations were measured at some of the cross-section profiles
318 when topo-bathymetric surveys were performed. The vertical accuracy of measured
319 stage is of 15 mm + 1 ppm, characteristic to GPS RTK/kinetic tracking method (with
320 a L1/L2 Topcon GNSS GR-3 receiver). At the GSs the overall accuracy of stage
321 records by staff gauge reading is about 5 mm or 0.2% of depth, whichever is greater
322 (Sauer and Turnipseed 2010). Comparison between measured and computed depth,
323 performed with the same discharge value, may be seen in Fig. 9.
324

325
326 Fig. 9 Relative errors in depth (measured vs. computed) in surveyed cross-sections along the
327 study reach
328
329 No velocity and discharge measurements with current meter were performed along
330 the study reach, since the accuracy of these would have been much lower than of
331 stage measurements, (errors up to 20%, according to Turnipseed and Sauer (2010)).
332 Therefore, for the calibration process it was only applied the commonly used

11
333 performance criterion of one dimensional hydraulic models in river hydraulics:
334 minimization of the difference between observed and computed stage / depth values
335 (Vidal et al. 2007, U.S. Army Corps of Engineers 1993).
336
337 Abstracted flow
338
339 When SHP plants operate, the flow is divided between the hydropower scheme and
340 the PFDR, in the channel flowing only a residual flow. A downwards shifted FDC
341 was obtained at Campina (Fig. 10), due to the abstraction at the intake of the
342 discharge for the operation of SHP with one turbine (operation time of 230 days /
343 year, corresponding to a probability of 65%) or both turbines (operation time of
344 130 days / year corresponding to a probability of 36%), each having a design
345 discharge of 4 m3/s. The greatest relative flow abstraction takes place during
346 moderate to high flows, whereas at flows lower than minimum turbine discharge for
347 efficient running, the last turbine unit has to stop.
348

349
350 Fig. 10 Potentially modified Flow Duration Curve at Campina GS for SHPs operation with
351 one and two turbines
352
353 As an example on how SHP operation particularily affects mid-range and low flows
354 (during winter and fall months), in Fig. 11 are shown at Campina GS, for the year
355 2012, the residual flow in the channel, Qresidual after the abstraction of Qabstracted for the
356 SHP, and the natural flow, Qnat as the sum of previous two values. One may see in
357 Fig. 12 that residual daily mean discharge values, registered (from the rating curve)
358 and measured (with a current meter, through the velocity-area method) at Campina
359 GS, in 2012, after the commissioning of upstream SHPs, have a low variability, since
360 discharge values rise only during the March - June floods.
361

12
362
363 Fig. 11 Monthly abstracted flow at the SHP (Qabstracted SHP), residual (Qresidual) and natural flow
364 (Qnat.) along the PFDR at Campina GS for the year 2012
365

366
367 Fig. 12 Daily mean values of residual flow (Qresidual) at Campina G.S., for the year 2012, after
368 the commissioning of the SHPs
369
370 Considering Qmean as the natural discharge flowing along the entire Prahova River
371 study river reach, a proposed flow split between penstock pipe (abstracted flow,
372 Qabstracted ) and channel (residual flow, Qresidual) was firstly computed for three
373 scenarios:
374 a) Without the operation of SHPs (natural regime). In the channel flows Qnatural =
375 Qmean.

13
376 b) SHPs operates at partial load, with only one turbine (Qdesign 1 turbine). Abstraction of
377 the design discharge for only one turbine at all of the SHPs. As a consequence,
378 the residual flow
379 Qresidual  Qmean  Qdesign 1turbine (2)
380 flows into the river channel; and
381 c) With operation of SHP at nearly full load. Maximum allowable abstracted flow
382 rate (Qmax. turb.) at all SHPs, so that only the minimum flow, Qmin is left as residual
383 flow in the PFDR (as the “worst case scenario” from environmental point of
384 view). The following algorithm was used to compute abstracted flow:
If (Qmean  Qmax. turb . )  Qmin
then Qresidual channel  Qmean  Qmax. turb .
385 and Qabstracted  Qmax.turb . (3)
Else Qresidual channel  Qmin
Qabstracted  Qmean  Qresidual channel

386 Computed flow values for abstracted and residual flow with relations (2) and (3) for
387 the considered three scenarios are shown in Table 2 and Fig. 13

388
389 Fig. 13. Channel residual flow, Qresidual, for different scenarios of SHPs operation: a) Qmean (no
390 SHP), b) Qresidual = Qmean - Qdesign 1 turbine (only 1 turbine) and c) Qresidual = Qmin. (all turbines)
391

14
392
393 Table 2 Computed abstracted and residual flow values
Scenario no SHP 1 turb. operation "Hands off" flow
DS SHP name Q max (a) (b) (c)
Dist. turb. Q min Q mean Q abstr Q resid Q abstr Q resid
(km) (m3/s) (m3/s) (m3/s) (m3/s) (m3/s) 3
(m /s) (m3/s)
36.69 Sinaia 0 intake (dam) 6.0
35.46 Sinaia 0 SHP 5.2 0.9 6.0 1.3 4.70 5.10 0.9
33.96 Sinaia 1 SHP 5.2 0.9 6.0 1.3 4.70 5.10 0.9
33.59 Sinaia 2 intake (dam) 6 1 6.9 2 4.90 5.90 1
32.80 Sinaia 2 SHP 6 1 6.9 2 4.90 5.90 1
31.61 Sinaia 3 SHP 6 1 6.9 2 4.90 5.90 1
30.97 Sinaia intake for Posada 7 1.1 7.15 3 4.15 6.05 1.1
394
395 RESULTS AND DISCUSSIONS
396
397 The following computation results are shown in comparison, between two residual
398 flow scenarios in the channel: (a) annual mean discharge, Qmean, and (c), estimated
399 minimum discharge, Qmin.
400 Computed average velocity along the PFDR corresponding to Qmean (scenario a)),
401 resulted to be of 1.77 m/s (Fig. 14). This value drops about 40% to an average value
402 of 1.1 m/s, for scenario c). There is only one downstream location where this value
403 drops below 0.5 m/s.
404

405
406 Fig. 14 Stream velocities for Qmean (no SHP operation, scenario a)) and for Qmin (with SHP
407 operation, scenario c))
408
409 Maximum channel depths, Dmax, also decrease at about half between scenario a) and
410 c), from an average value of 0.53 m to an average of 0.26 m (Fig. 15). There are few
411 locations, particularily between km 23 and km 26, where channel depth seems to drop
412 lower than the 20cm limit of fish habitat conditions. Further investigations are needed
413 for this area.
414

15
415
416 Fig. 15 Maximum stream depth for Qmean (no SHP operation, scenario a)) and for Qmin (with
417 SHP operation, scenario c))
418

419
420 Fig. 16 Froude no. values with computed velocity (V) and mean depth (D), for Qmean (no SHP
421 operation, scenario a)) and for Qmin (with SHP operation, scenario c)) in the channel
422
423 In Fig. 16 are shown Froude number values along the study reach, calculated with the
424 relation
V
425 Fr  (4)
gD

426 where V and D are the computed velocity and hydraulic depth respectively, in each
427 cross-section. Froude number is known to be a good physical descriptor of fish
428 community patterns between different stream reaches and river hydraulics
429 (Lamouroux and Cattanéo 2006). As one may see, the average value of this parameter
430 (and corresponding regime) remains fairly constant with discharge, (0.8, compared to
431 0.71), the most important differences between the two scenarios being along the

16
432 Carpathian reach (km 25-36). At Campina GS (km 2), average computed values are
433 of about 0.6.
434 In Fig. 17 are represented, for comparison, measured values of depth and velocity at
435 Campina GS, for the year 2012, after the commissioning of SHPs. One may see
436 Froude number is in the range 0.26-1.05, with a mean value of 0.57. The values are
437 similar to the ones obtained with the same formula for mountain, gravel-bed streams
438 by other authors: 0.2-1.2, with a mean value of 0.6 (Ancey et. al. 2014), 0.1-1 (Jowett
439 2010) and 0.1-1.2 (Steele and Lorang 2009). On the same plot in Fig. 17 are
440 represented computed velocity, depth and Froude number for the discharge registered
441 the day when topo-bathymetric survey was performed, in the same cross-section (at
442 Campina GS). As it may be seen from this plot, transitions from subcritical to
443 supercritical regime in this cross-section arises only during spring flood (in April).
444

445
446 Fig. 17 Measured and computed values of velocity (V), mean depth (D) and Froude no. at
447 Campina GS, in 2012
448
449 Average flow area decreases more than four times between the two considered
450 scenarios (Fig. 18). This important change could affect water temperature during
451 winters (Yapa and Shen 1984) and summer months, with ecological consequences.
452 High values of flow area at km 36 and km 5 correspond to the Sinaia 0 reservoir, and
453 the river exit of the Sub-Carpathians (where the channel broadens), respectively.
454 There are few particular locations (between km 10 and km 30) where cross-sectional
455 areas are even lower than 0.5 m2.
456

17
457
458 Fig. 18 Channel wetted area for Qmean (no SHP operation, scenario a)) and for Qmin (with SHP
459 operation, scenario c))
460
461 Average bed shear stress, , is an important parameter to analyse potential sediment
462 bedload transport, driven by the hydraulic parameters of the flow (Parker 2008). It
463 was computed with the following relation, valid for uniform flow (either subcritical
464 or supercritical):
465   gRh S f (5)
466 in which  is the water density, g is the acceleration of gravity, R = A/P is the cross-
467 section hydraulic radius, and Sf – the friction slope.
468

469
470 Fig. 19 Bed shear stress for Qmean (no SHP operation, scenario a)) and for Qmin (with SHP
471 operation, scenario c))
472
473 Fig. 19 shows an overall two times reduction in computed bed shear stress values
474 between the two scenarios. Peak values can be explained by sudden changes in slope,
475 triggering regime to pass through critical one. These changes are encountered at drop
476 structures, such as rapid hydraulic structures, energy dissipators or check-dams. For
477 example, at km 26 (near Posada SHP) there is a long stepped energy dissipator, its
478 slope sudden increase introducing a high value of bed shear stress. Also, from the

18
479 same Fig. 19 it may be seen that along the Posada Defile and the Sinaia intake for
480 Posada SHP (km 24-26), the values of shear stresses are high.
481 To analyse the onset of bedload transport for the median diameter of the sediment
482 particle distribution (d50) of the surface layer and substrate, nondimensional bed shear
483 stress (Shields parameter), *

484   (6)
   s   1gd 50
485 was compared with a threshold - critical value, the nondimensional bed shear stress,
486 *cr, computed with the Parker formula (Parker 2008) for gravel-type sediment
487 particles as a function of particle Reynolds no., Rep, associated to the median
488 diameter particle, d50:
489
( 7.7 Re p0.6 )  s   1gd 50 d 50
490  cr  0.22 Re p0.6  0.06  10 , Re p  . (7)

491
492 In equations (6) and (7)  is the kinematic viscosity of water, s is the density of
493 sediment particle (taken as 2650 kg/m3 for quartz) and  s   1 is the submerged
494 specific gravity of the sediment.
495

496
497 Fig. 20 Onset of bedload transport (median particle of the surface layer, d50 = 28mm) for
498 Qmean (no SHP operation, scenario a)) and for Qmin (with SHP operation, scenario c))
499
500 Inception of movement was approximated by the condition * > *cr,. In Fig. 20
501 computation results are presented for the two same comparative scenarios (a) and (c)
502 and the meadian particle of the surface layer with d50 = 28 mm. One may see that
503 bedload transport corresponding to the coarse gravel class of particles generally
504 ceazes to occur at minimum flow Qmin in the river channel (scenario c)).
505 Computations with other particle sizes showed the largest transported particle by Qmin
506 has a diameter of 10 mm (fine gravel). This means that for low flows, only sand and
507 very fine gravel particles may be transported over the PFDR, thus covering areas
508 where coarse gravel, cobbles and boulders were deposited over the surface layer by

19
509 previous floods (Nistoran et al. 2007). As a consequence, fish habitat requesting
510 larger sediment particles might suffer. However, the equilibrium may be restored and
511 larger particles may be transported again during subsequent high flow periods
512 (Rodrigues et. al. 2006, Rodrigues et. al. 2014).
513 Because of that, specialists in the field of RoR SHPs recommend PFDR to be
514 monitored for a couple of years after commissioning to ensure sedimentation with
515 fine particles is not occuring. If sedimentation is observed to be a problem, then a
516 “turbine-off” protocol should be incorporated into the operating regulation to allow
517 the gravels to be regularely flushed.
518
519 CONCLUSIONS
520
521 Flow diversion for RoR SHPs affects the natural streamflow, particularily for middle
522 range and low discharge values, reducing the natural flow variability. Changes
523 induced to the hydrodynamic parameters along the PFDR were compared and
524 analysed using a 1D hydraulic model. Two limiting scenarios were considered for the
525 residual flow in the channel: (a) annual mean discharge, Qmean, (average natural
526 situation without the operation of SHPs) and (c), estimated minimum discharge, Qmin,
527 (situation encountered during intermediate and low flow periods when operating the
528 SHPs at maximum allowable discharge, Qmax turb.).
529 On average, results along the study reach show a 50% decrease in water depth and
530 shear stress, 40% in velocity, 75% in flow area and no significant decrease of Froude
531 numbers for scenario (c) as compared with scenario (a). However, lowest values of
532 velocity and depth are still above minimum required fish habitat conditions (exept for
533 very few locations, where further investigations are needed).
534 An analysis was performed to check the onset of bedload transport during low flow
535 periods, at Qmean and Qmin in the channel. Nondimensional bed shear stress (Shields
536 parameter) was compared with the critical value for the inception of motion.
537 Computations showed that only sand and very fine gravel particles (with d50 < 10 mm)
538 may be transported over the PFDR, covering coarse gravel, cobbles and boulders
539 already deposited over the surface layer. This could potentially affect fish habitat in
540 the absence of floods or hydropower shut-offs. If necessary, operation rules should
541 also take into account fish requirements during the spawning and hatching season.
542 Present paper shows the importance of hydraulic modeling along the potentially
543 depleted reaches. This kind of studies should be performed prior to the design and
544 construction of RoR SHP, to verify the compliance of minimum habitat requirements
545 of encountered species in the river channel.
546
547 Nomenclature
548
549 D = hydraulic depth (area divided by top width) (m)
550 Dmax = maximum depth in cross-section (m)
551 Dmeas = measured depth (m)
552 Dcomp. = computed depth (m)
553 DTM = digital terrain model

20
554 d50 = median diameter particle of the sediment distribution (with 50% probability)
555 (mm)
556 FDC = mean daily flow duration curve
557 Fr = Froude number
558 GS = gauging station for river discharge/stage values
559 SHP = small hydropower plant
560 PFDR = potentially water depleted reach
561 Qmean =multiannual mean flow (m3/s)
562 Qmin. = minimum flow in the stream (m3/s)
563 Qe = environmental (ecological) flow (m3/s)
564 Qm.m.= monthly minimum flow (m3/s)
565 Q95 = minimum monthly flow that corresponds to 95% probability of exceedance on
566 the flow duration curve (m3/s)
567 Qabstr. = abstracted flow from the river to the penstock pipes (m3/s)
568 Qnat. = natural flow in the river channel taken as Qmean (m3/s)
569 Qresidual = residual flow in the river channel, after the abstraction of Qabstr. to the SHP
570 penstock pipes (m3/s)
571 R = hydraulic radius (m)
572 Rep = Reynolds number associated to mean diameter sediment particle
573 RoR = run-of-river
574 Sf = friction slope
575 TIN = Triangular Irregular Network
576 W.S. elev = water surface elevation
577  = bed shear stress (N/m2)
578 * = nondimensional bed shear stress (Shields parameter) (N/m2)
579 *cr = value of nondimensional bed shear stress critical for motion inception (N/m2)
580
581 REFERENCES
582 Ancey, C., Bohorquez, P., Bardou E., (2014). “Sediment transport in mountain rivers.”,
583 ERCOFTAC Bulletin, 100, Sept., 37-52.
584 Anderson, E., Freeman, M. and Pringle, C. (2006). “Ecological consequences of Hydropower
585 Development in Central America: Impacts of Small Dams and Water Diversion on
586 Neotropical Stream Fish Assemblages.” River Res. Appl., 22(4), 397- 411.
587 Anderson, D., Moggridge H., Warren, P., Shucksmith, J. (2015). “The impacts of Run-of-
588 river hydropower on the physical and ecological conditions of rivers.” Water and
589 Environment Journal, 29, 268-276.
590 Angermeier, P.L., Davideanu, G. (2004). “Using fish communities to assess streams in
591 Romania: initial development of an index of biotic integrity.” Hidrobiologia, 511, 65-78.
592 Arlinghaus, R., Wolter, C. (2003). “Amplitude of ecological potential: chub Leuciscus
593 cephalus (L.) spawning in an artificial lowland canal.” J. Appl. Ichthyol, 19, 52–54.
594 Armaş, I., Gogoaşe Nistoran, D.E., Osaci-Costache, G., Braşoveanu L. (2012). “Morpho-
595 hydrodynamic evolution patterns of Subcarpathian Prahova River (Romania).” Catena,
596 100, 83-99.
597 Arthington, A.H. and Zalucki, J.M. (Ed.s) (1998). “Comparative evaluation of Environmental
598 Flow Assessment Techniques: Review of Methods”, Land and Water Resources Research
599 and Development Corporation (LWRRDC) Occasional Paper 27/98, Centre for Catchment
600 and In-Stream Research, Griffith University, Canberra, Australia.

21
601 Benejam, L., Saura-Mas, S., Bardina M., Soia C., Munne A., Berthou-Garcia, E. (2016).
602 “Ecological impacts of small hydropower plants on headwater stream fish: from
603 individual to community effects.” Ecology of freshwater, 25, 295-306.
604 Bruno, G.S. and Fried, L. (2008). “Focus on Small Hydro.” Renewable Energy Focus, 9, 54-
605 57.
606 Csiki, S., Rhoads, B. (2010). “Hydraulic and Geomorphological Effects of Run-of-River
607 Dams.” Prog. Phys. Geog., 34 (6), 755-780.
608 Demirabas, A. (2007). “Focus on the World: Status and Future of Hydropower.” Energy
609 Source, Part B: Economics, Planning, and Policy; 2 (3), 237-242.
610 Environment Agency (2009). “The environmental assessment of proposed low head hydro
611 power developments”. in Good Practice Guidelines Annex to the Environment Agency
612 Hydropower Handbook; Rotterdam, UK.
613 ESHA - European Small Hydropower Association, Thematic network on small hydroelectric
614 plants, Environmental group, (2005). “Reserved flow – short critical review of the
615 methods of calculation”, <http://www.surna.no/web/NettskoleRessurs.axd?id=4aa21e26-
616 93fc-474a-8ee2-c2aff3766fab> (July 21, 2017).
617 European Commission (2015). “Ecological flows in the implementation of the Water
618 Framework Directive. Guidance document no. 31”, Technical Report, Luxembourg:
619 Office for Official Publications of the European Communities, European Union
620 European Parliament and Council of The European Union (2000). “Directive no. 2000/60/EC
621 establishing a framework for Community action in the field of water policy”, Official
622 Journal of the European Communities, L 327, Volume 43, 22.12.2000.
623 European Parliament and Council of The European Union (2009). “Directive 2009/28/EC on
624 the promotion of the use of energy from renewable sources and amending and
625 subsequently repealing Directives 2001/77/EC and 2003/30/EC”, Official Journal of the
626 European Union, L 140, Volume 52, 5.6.2009.
627 Fraenkel, P., Paish, O., Bokalders, V., Harvey, A., Brown A., Edwards, R. (1991).
628 Hydropower: a guide for development workers, IT Publications Ltd., London, U.K.
629 Gogoaşe Nistoran, D.E., Ionescu C.S, Pătru G., Armaş I., Grigorie Omrani Ş. (2017). “One
630 dimensional sediment transport model to assess channel changes along Olteniţa-Călăraşi
631 reach of Danube River, Romania.” Energy Procedia, 112, 67-74.
632 Gopal, B. (Ed.) (2013), “Environmental flows; An introduction for Water Resource
633 Managers”, National Institute of Ecology, New Delhi, India
634 http://www.aquaticecosystems.org/library/environmental-flows/ (July 21, 2017)
635 HEC - Hydrologic Engineering Center (2010). “River Analysis System.”, Version 4.1, User’s
636 manual, Hydraulic reference manual, U.S. Army Corps of Engineers, Institute for water
637 resources, Davis, CA, USA.
638 HEC - Hydrologic Engineering Center (2011). “Geo-RAS GIS tools for support of HEC-RAS
639 using ArcGIS”, Version 4.3.93, User’s manual, Hydraulic reference manual, U.S. Army
640 Corps of Engineers, Institute for water resources, Davis, CA, USA.
641 Huckstorf, V., Lewin W.-C., Wolter, C. (2008), “Environmental flow methodologies to
642 protect resources in human-modified large lowland rivers.”, River Research and
643 Application, 24, 519-527.
644 Ingram, A.; Ibbotson, A.; Gallagher, M. (1999). “The Ecology and management of European
645 Grayling Thymallus thymallus (Linnaeus) Interim Report.” Institute of Freshwater
646 Ecology, Center for Ecology and Hydrology, River Laboratory, Dorset, UK,
647 <http://aquaticcommons.org/8072/> (July 21, 2017).
648 Jowett, I. G. (2010). “A method for objectively identifying pool, run, and riffle habitats from
649 physical measurements.” New Zealand Journal of Marine and Freshwater Research,
650 27, 241-248

22
651 Kaunda, C. S., Kimambo, C. Z. and Nielsen, T. K. (2012). “Hydropower in the Context of
652 Sustainable Energy Supply: A Review of Technologies and Challenges.” ISRN Renewable
653 Energy, Volume 2012, Article ID 730631.
654 Kondolf, G.M., Wolman M.G. (1993). “The size of salmonid spawning gravel.” Water
655 Resources Research, 29, 2275-2285.
656 Lamouroux, N., and Cattanéo, F. (2006.). “Fish assemblages and stream hydraulics:
657 consistent relations across spatial scales and regions.” River Res. Applic., 22: 727-
658 736
659 Latapie, A., Camenen, B., Rodrigues S., Paquier, A., Bouchard, J.P., Moatăr, F. (2014).
660 “Assessing channel response of a long river influenced by human disturbance.” Catena,
661 121, 1-12.
662 Leopold, L.B., Wolman, M.G. (1957). “River channel patterns: braided, meandering and
663 straight.” US Geological Survey Professional Paper, 282-B US Government Printing
664 Office, Washington.
665 Leopold, L.B., Wolman, M., Miller, J. (1964). Fluvial Processes in Geomporhology, W.H.
666 Freeman, San Francisco, CA.
667 Melcher, A.H., Schmutz, S. (2010). “The importance of structural features for spawning
668 habitat of nase Chondrostoma nasus (L.) and barbel Barbus barbus (L.) in a pre-Alpine
669 river.”, River Syst.,19(1), 33-42.
670 Nistoran D.E., Moatăr F., Rodrigues S. (2007). “Hydraulic modeling of an anabranched river
671 reach: riffle-pool reversal and maintenance mechanism during a flood event.” Proc., 32nd
672 Congress of IAHR, Harmonizing the Demands of Art and Nature in Hydraulics, Venice.
673 Otuagoma, S. O., Ogujor E.A., Kuale P.A. (2016). “Evaluation of small hydropower potential
674 of river Ethiope using the RETScreen Software.” Intl. J. of Engineering Research, 5(1), 1-
675 5.
676 Parker, G. (2008). “Transport of Gravel and Sediment Mixtures” (Chapter 3) in García M.
677 H., (Ed.), “Sedimentation Engineering. Processes, management, modeling and practice.”,
678 ASCE Manual and Reports on Engineering Practice, 110, 165-251.
679 Pascu, M. (1999). “Research on the influence of radical regulation works in river channels on
680 the stability of construction works and on the environment – Prahova River basin.”
681 Ph.D. Thesis, Technical University of Civil Engineering, Faculty of Hydrotechnics,
682 Bucharest, Romania.
683 Punys, P., Dumbrauskas, A., Kasiulis, E., Vyčienė, G., and Šilinis, L. (2015). “Flow Regime
684 Changes: From Impounding a Temperate Lowland River to Small Hydropower
685 Operations.” Energies, 8, 7478-7501.
686 Rădoane, M., Rădoane, N., Dumitriu, D., Miclăuş, C. (2008). “Downstream variation in bed
687 sediment size along the East Carpathian rivers: evidence of the role of sediment
688 sources.” Earth Surf. Process. Landforms, 33(5), 674-694.
689 Robson, A., Cowx, I. and Harvey, J. (2011). “Impact of run-of-river hydro schemes on fish
690 populations: Phase 1-Literature review”. SNIFFER - Scotland and Northern Ireland
691 Forum for Environmental Research - (SNIFER), Edinburgh, Scotland, UK,
692 ”<https://www.sniffer.org.uk/Handlers/Download.ashx?IDMF=7cd16fc7-4d46-4ae9-
693 a6b8-b9e191656477> (May 25, 2017).
694 Rodrigues, S., Breheret, J.-G., Macaire, J.-J., Moatăr, F., Nistoran, D., Juge, P. (2006). “Flow
695 and sediment dynamics in the vegetated secondary channels of an anabranching river: the
696 Loire river (France).” Sedimentary Geology, 186, 89-109.
697 Rodrigues, S., Mosselman, El., Claude, N., Juge, P. (2014). “Alternate bars in a sandy gravel
698 bed river: generation, migration and interactions with superimposed dunes.” Earth Surface
699 Processes and Landforms, 40(5), 610-628.

23
700 Romanian Government (2007 a). “Government Decision no. 1069 on the National Strategy
701 for Energy in Romania for the 2007-2020 period” (In Romanian), Romanian Official
702 Journal, no. 781, 19.11.2007.
703 Romanian Government (2007 b). “Emergency Order no. 57/2007 on juridical regime of
704 protected areas, natural habitat conservation, wild flora and fauna.” (in Romanian),
705 Romanian Official Journal, no. 442, 29.06.2007
706 Romanian Ministry of Environment and Forests (2012). “Order no. 799 for approving the
707 content of the Technical documentation required to obtain the water consent and water
708 permit”, Romanian Official Journal, no. 151, 7.3.2012.
709 Romanian Parliament (1996), “Water Law” (In Romanian), Romanian Official Journal, no.
710 244, 8.10.1996
711 Romanian Parliament (2008). “Law establishing the promotion system for the production of
712 energy from renewable energy sources” (In Romanian), Romanian Official Journal, no.
713 743, 3.11.2008.
714 Sauer, V. B. and Turnipseed, D.P. 2010. “Stage measurement at gaging stations: Techniques
715 and Methods, book 3, chap. A7.” U.S. Department of the Interior, U.S. Geological Survey,
716 Reston, VI, USA <http://pubs.usgs.gov/tm/tm3-a7/pdf/tm3-a7.pdf> (May 25, 2017)
717 Stematiu, D. (2008). “Hydropower developments” (in Romanian), Conspress, Bucharest,
718 Romania.
719 Toroimac, G.I., Dobre, R., Grecu, F., Zaharia. L. (2010). “Evolution 2D de la bande active de
720 la Haute Prahova (Roumanie) durant les 150 dernières années.” Géomorphologie: relief,
721 processus, environnement, 3, 275-286.
722 Turnipseed, D.P and Sauer, V. B. (2010), “Discharge measurement at Gaging Stations:
723 Techniques and Methods, book 3, chap. A8.” U.S. Department of the Interior, U.S.
724 Geological Survey, Reston, VI, USA <https://pubs.usgs.gov/tm/tm3-a8/pdf/tm3-a8.pdf>
725 (May 25, 2017).
726 U.S. Army Corps of Engineers (1993). “River Hydraulics” Engineering Manual no. 1110-2-
727 1416, USA,
728 <http://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1
729 110-2-1416.pdf?ver=2013-09-04-070758-670> (May 2017).
730 Vehanen, T., Huusko, A., Yrjänä, T., Lahti, M., Mäki-Petäys, A. (2003). “Habitat preference
731 by grayling (Thymallus thymallus) in an artificially modified, hydropeaking riverbed: a
732 contribution to understand the effectiveness of habitat enhancement measures.” J. Appl.
733 Ichthyol, 19, 15–20.
734 Vidal, J.-P., Moisan, S., Faure, J.-B., Dartus, D. (2007), “River model calibration, from
735 guidelines to operational support tools”, Environmental Modelling and Software, 22,
736 1628-1640.
737 Wohl, E. (2006). “Human impacts to mountain streams.” Geomorphology, 79, 217-278.
738 Yapa, P. and Shen, H. (1984). “Effect of Ice Cover on Hydropower Production" J. Energy
739 Eng., 110:3 (231), 231-234.
740

24

You might also like