1* 1 2 1 Raster based modelling of watershed erosion and sedimentation in an Ravazzani , G., Rulli , M.C., Groppelli , B., Bocchiola , D., (1) Politecnico di Milano, 2 1 1 Alpine basin of Northern Italy Colombo , F., Mancini , M., Rosso , R. (*) corresponding author: [email protected] (2) ENVIRON, Italy

ABSTRACT: Soil erosion and sedimentation assessment is fundamental in both engineering practice and hydrologic research. A MODEL DESCRIPTION wide range of mathematical models have been developed during years; they can be classified in two categories: a) models mainly METEO SURFACE RUNOFF SEDIMENT DETACHMENT Meteo Data oriented to hillslope surface erosion that concentrate attention to rill precipitation EXTENDED SCS-CN Splash erosion = 2 ( ) Dir .0 0138 i K SF Cslr and interrill processes, and b) models developed for simulating in- Spatial SF = 05.1 − 8.0 EXP (− sin4 θ ) Interpolation 2 = ⋅ −ε Thiessen − St S1 (1 t ) stream phenomena mainly related to bed-load transport. In order to (P Ia ) i= rainfall rate R = lorda Splash − + K= soil erodibility(USLE) predict erosion and sedimentation at the river basin scale, all Erosion P Ia S Cslr = land use soil loss ratio (USLE) b= slope sediment related processes on hillslope and river channel need to be θ −θ S Snow Pack DEM t res simulated in an integrated model. Moreover, in poorly gauged ε = Channel erosion temperature t θ −θ ε sat res = α β mountainous river basins, where available measurements and data Dr Q KSC slr Soil Vegetation Glaciers Least cost Least are generally limited, models need to be kept as simple as possible in algorithm Reference plane Q= discharge order to limit the number of parameters. Flow Paths S= slope Vertical thermal α= calibration parameter Parameters Soil Moisture β= calibration parameter The erosion and sedimentation model (ERODE) developed in this gradient S map before a Update flood event study combines raster USLE parameters to the spatially distributed Hillslope and Channel Sun, H., P. S. Cornish, and T. M. Daniell (2002), physically based hydrological model FEST-WB processes with Network Digital Elevation Contour-based digital elevation modeling of watershed Definition erosion and sedimentation: Erosion and sedimentation Hillslope Channel Model estimation tool (EROSET), Water Resour. Res., 38(11), minimal data requirement. In particular, the model takes in account 1233. rain splash detachment in overland areas and shear stress erosion in channel areas. Sediment transport and sedimentation are evaluated Deep Surface Shear Stress Runon SNOW ROUTING SEDIMENT TRANSPORT along the river network by calculating the transport capacity trough Drainage runoff Erosion SNOW ACCUMULATION Sediment routing the unit stream power theory. α  P = αP L Variable celerity Route:  =τ ⋅ W = sediment storage Route: linear neve neve + pioggia pioggia W Q The model was applied to an Alpine river basin in Northern Italy Muskingum- P = ()1−α P Muskingum- Q = sediment discharge reservoir S Cunge τ = travel time Sediment Cunge ∆s and it was validated against measurements of suspended solids and Initial 1 τ = ∆s= channel length transport and  v0 condition + + v = normal flow velocity deposition  Q j 1 = C ⋅Q j 1 + C ⋅ Q j + C ⋅Q j 0 total sediment transport at the basin scale. The results show that the 0 if T ≤ T i+1 1 i 2 i 3 i+1  a low α = ≥ ∆ − ⋅ ⋅ε p 1 if Ta Tup t 2 k Tinf Tsup T(°C) C = model is able to correctly simulate storm and within-storm erosion Reservoir  T − T 1 ⋅⋅ ()− ε + ∆ a low < < 2 k 1 t routing  if Tlow Ta Tup Transport capacity - deposition Outflow Outflow T − T ∆ + ⋅ ⋅ε and deposition.  up low = t 2 k Hydrograph Sedimentograph C2 SNOW MELT 2 k ⋅⋅ ()1− ε + ∆t 25.0 t  Q  S 375.1 ⋅ ⋅ − ε − ∆ P =   A C = transport capacity = 2 k (1 ) t  R  n 75.0 t +1j C3 P = unit stream power (j+1) ∆∆∆t Qi + ⋅⋅ ε− + ∆ Q 1j 2 k (1 ) t +1i R = hydraulic radius ∆ 1  Q  ∆∆∆t x =ε ⋅  −   P − P  ω= settling velocity degree-day + = 1 cr 1j k   = γ + ε q + ⋅ω⋅ ∆⋅ log Ct log   1i L 1i ω 2  B S0 x  ω n= Manning roughness ∆∆∆ j   j t Q +1i Qj ω A = wetted area M = C T( −T ) i i,j (t) = 5/3 Vi,j (t) ACKNOWLEDGEMENT: This work carried out under the s m a b ∆∆∆s

2 0.75   0.25 Moore, I. and Burch, G. (1986) Physical Yang, C.T. (1 973). "Incipient Motion and umbrella of the Characterization of Lake Maggiore and Toce River    3    rf 0.5 Q Sediment Transport," Journal of the V = k   ⋅i  ⋅  basis of the length-slope factor in the s  +    Universal Soil Loss Equation. Soil Society Hydraulic Division,   rf 2    rf  project, funded by Syndial SpA, which also supported the 0i ∆∆∆s (i+1) ∆∆∆s s   of America Journal, 50, 1294 – 1298.. ASCE, vol. 99, no. HY 10, pp. 1679-1704. characterization activities. Thanks to ARPA Piemonte for providing temperature and precipitation data from their stations. CASE STUDY SIMULATION RESULTS

River Toce Basin - Italy

10,000 0 #S N N N 10,000 0 W E W E FORMAZZA BRUGGI W E #S The subject area is the river Toce basin, a typical alpine ALPE DEVERO S #S S 1,000 #S Soil Erodibility S cover factor 1,000 10 ALPE VEGLIA basin with a total drainage area of about 1700 km², with #S 10 0 - 0.2 #S 0 - 0.01 100 nearly 32% of the total area above 2000 m a.s.l.. It is #S #S LARECCHIO 0.2 - 0.4 #S 0.01 - 0.02 located in the north of the region in Italy (Fig. LAGO PAIONE #S 100 20 10 0.02 - 0.03 0.4 - 0.6 20 DRUOGNO 1). Climate conditions are typically humid, characterized PIZZANCO #S #S 0.6 - 0.8 #S 0.03 - 0.04 1 by higher precipitations in autumn and spring. The annual ALPE CHEGGIO #S 0.8 - 1 0.04 - 0.05 30 average precipitation exceeds 2000 mm. Climatic 10 30 0 FOMARCO

PASSO DEL MORO (mm/h) Precipitation characteristics, together with morphology and soil #S ANZINO #S #S #S CANDOGLIA (mm/h) Precipitation #S #S observed texture, frequently induced flood events in the past years. 0 #S LAGO DELLE LOCCE simulated simulated 40

SAMBUGHETTO (mg/l) concentration Sediment Most of the presented results are related to the Anza #S 1 40 0 MOTTARONE observed precipitation #S#S 0 10 20 30 Kilometers river basin, a tributary of the river Toce, with a total MOTTARONE-BAITA CAI 0 10 20 30 Kilometers (mg/l) consentration Sediment precipitation drainage area of about 261 km². #S 0 50 0 10 20 Kilometers 5/2/2010 7/1/2010 8/30/2010 10/29/2010 Available digital cartographic data include: the Digital 0 50 Elevation Model (DEM) available in raster format at 100 3/8/2009 3/10/2009 3/12/2009 m x 100 m resolution; CORINE land cover maps (CEC, The Toce watershed extracted from the Validation of suspended solid simulation: comparison Validation of suspended solid simulation: comparison 1994, EEA, 2000) updated in the year 2000 available in digital elevation model showing locations Soil erodibility factor (kgh/ N/m 2) Cover and management factor between continuous turbidity measurements and simulation between sampled TSS and simulation at Candoglia vector format; pedologic characteristics for soils available of the 25 rain gauges at Candoglia in vector format.

4.00E+09

3.50E+09

3.00E+09

disponibili 2.50E+09 Tratti omogenei observed 2.00E+09 simulated 1.50E+09 Sediment (kg) Sediment

1.00E+09

5.00E+08

0.00E+00 2000 2007 ->2009 2007 1982 -> 1989 Available bed sediment Position of samples for 2001 -> 2006 1982 -> 2009

July 1996 -> 1999 July -> 1996 Simulated annual sediment inflow into lake laboratory analysis Total Suspended Solid Continuous turbidity sediment deposition rate -> June 1990 1996 sample acquisition measurement: 20 Days in estimation in the bay Validation of total (suspended + bedload) sediment April 2009 simulation: comparison between measured sediment deposition in the bay and simulation at basin outlet