Role of hydrological model uncertainties in climate change impact - PowerPoint PPT Presentation

Role of hydrological model uncertainties in climate change impact studies Satish Bastola, Conor Murphy, and John Sweeney ICARUS, NUIM Ireland HydroPredict’ 2010: 20 -23 September 2010 Prague, Czech Republic

Contents • Introduction • Method – Account for hydrological model uncertainty – quantify uncertainty in impact studies • Results • Conclusion

Uncertainty that cascade through a climate change impact assessment • Projected changes in future climate are inherently uncertain Emission Scenarios: Economic activity, population growth, Technology Response of climate model to emissions Impact models (simple eqn, spa/tem aggr, calib,plausable • Considerable work have focused on Emission and GCM uncertainty but have mostly neglected uncertainties in impact models

Objective Examine the role of model uncertainty (parameter and structural uncertainty) in climate change impact studies A. Account for hydrological model uncertainty (GLUE, BMA) B. Quantify uncertainties that cascade through the climate change impact assessment

Schematic: accounting for uncertainty in Hydrological model using GLUE GLUE method to account for Para & Str.Uncertainty in Hydrological models Simulators Model 1 L Parameter θ θ Likelihood L Threshold TH   ( I , ) Y f L Model 2 θ θ Streamflow L TH L Model k Time θ θ Time (2000-2100) pdf of parameter L Simulation TH GLUE: Generalized Likelihood Uncertainty Estimation Method (Beven and Binley, 1992) L : Likelihood; θ : Model parameters; TH: threshold of Likelihood GLUE has been extensively used (e.g. Freer et al., 1996; Freer et al., 2004; Montanari, 2005 and more)      2 2 ( | ) 1 ( / ) L i Y i obs

Bayesian Model averaging In BMA the predictive probability density function (PDF) of any quantity of interest is a weighted average of PDFs centred on the individual forecasts K     ( | ,.., , ) ( | , ) ( | ) p M M D p M D p M D 1 K k k  1 k Weight (Wk) Δ Posterior distribution of BMA prediction x p( Δ | Mk) M1 Mk x x p( Δ |M1) x Time (Tutorial on BMA: Hoeting et al., 1999)

Bayesian Model Averaging GLUE For each conditional PDF gamma distribution was selected  e      1 ( / ) Model 1   | M k   )   (         2 2 2 Model 2 / ; / W 1 ,W2..Wk, k k k k k b, C o      2 ; . M b M c k k k k o Model k Weight and variance parameter of BMA were estimated using DREAM of Vrugt et al (2008). For BMA n       2 2 ( ,.. | ,.. , ) log( ( | ) l w w w P M 1 1 1 1 k k  1 t     ( | ) .. ( | )) Time (2000-2100) w P M w P M 2 2 k k K     ( | ,.., , ) ( | , ) p M M D p M D W 1 K k k  1 k

Study Region & Data Irish National Meteorological Service • Office of Public works, ireland Six regional climate scenarios from Fealy and Sweeney, 2007 A2 B2 HadCM3 CCCMA CSIRO HadCM3 CCCMA CSIRO Republic of Ireland CCCma (CGCM2):canadian centre for climate modeling and analysis; CSIRO: Commonwealth Scientific and Industrial Research Organization; HadCM3:Hadley Centre Coupled Model, version 3

Hydrological model The Hymod, NAM, TANK, and TOP models describes the behaviour of each individual component in the hydrological cycle, at catchment level, by using a set of mathematical equations. HyMOD CQOF,TOF,TIF, TG TOP Model (Beven et al1995) Beven,1984 S  Beven and wood,1993) u z q v S Dt i d S (Beven,1991)   r z ( 1 ) E E TANK (Sugawara, 1995) a p NAM (DHI) S m a x r

Model uncertainty using GLUE/BMA Simulated flow:Daily (3yr)/seasonal(1971-1991) HYMOD_MOY W hymod , b, C o Range med Obs 150 prediction quantile (Cal (1971-1989):Hymod) BOYNE 350 Cumecs 300 100 Hymod 250 Streamflow (Cumecs) W Nam , b, C o 200 50 150 100 0 50 1 3 5 7 9 11 0 1/1/81 1/3/81 1/5/81 1/7/81 1/9/81 1/11/81 1/1/82 1/3/82 1/5/82 1/7/82 1/9/82 1/11/82 1/1/83 1/3/83 1/5/83 1/7/83 1/9/83 1/11/83 month (climatological axis) Range NAM_MOY med Time Obs prediction quantile (Cal (1971-1989):NAM) BOYNE 350 150 parameter & model 300 prediction quantile (Cal (1971-1989):Multimodel) BOYNE NAM Cumecs 250 100 350 Streamflow (Cumecs) structural uncertainty 200 300 150 50 250 Streamflow (Cumecs) 100 200 50 0 0 150 GLUE BMA 1/1/81 1/3/81 1/5/81 1/7/81 1/9/81 1/11/81 1/1/82 1/3/82 1/5/82 1/7/82 1/9/82 1/11/82 1/1/83 1/3/83 1/5/83 1/7/83 1/9/83 1/11/83 1 3 5 7 9 11 100 Time Range Tank_MOY month (climatological axis) 50 med prediction quantile (Cal (1971-1989):Tank) BOYNE Obs 0 350 1/1/81 1/4/81 1/7/81 1/10/81 1/1/82 1/4/82 1/7/82 1/10/82 1/1/83 1/4/83 1/7/83 1/10/83 300 150 250 Streamflow (Cumecs) Time Cumecs 200 100 Tank 150 100 50 50 0 1/1/81 1/3/81 1/5/81 1/7/81 1/9/81 1/11/81 1/1/82 1/3/82 1/5/82 1/7/82 1/9/82 1/11/82 1/1/83 1/3/83 1/5/83 1/7/83 1/9/83 1/11/83 0 1 3 5 7 9 11 Time prediction quantile (Cal (1971-1989):TOP BOYNE Range TOP_MOY month (climatological axis) W Tank , b, C o 180 med 160 140 Obs Streamflow (Cumecs) 120 BMA 150 100 80 Cumecs 60 100 40 Top 20 W TOP , b, C o 0 1/1/81 50 1/3/81 1/5/81 1/7/81 1/9/81 1/11/81 1/1/82 1/3/82 1/5/82 1/7/82 1/9/82 1/11/82 1/1/83 1/3/83 1/5/83 1/7/83 1/9/83 1/11/83 Time 0 1 3 5 7 9 11 Accounts for parameter month (climatological axis) uncertainty

Model calibration/validation Period Basin (Model) NSE (Median) CE PI (m3/s) Sn (Calib/Valid) Calib Valid Calib Valid Calib Valid 1 Moy (HYMOD) 0.77 0.66 0.68 0.56 30.50 33.01 2 1971-1990/1991- Moy (NAM) 0.72 0.63 0.58 0.52 25.69 27.66 2000 3 Moy (TANK) 0.80 0.69 0.80 0.77 40.88 44.55 4 Moy (TOP) 0.80 0.70 0.72 0.70 33.98 37.47 Ensemble Med 0.81 0.72 0.85 0.80 43.32 46.84 5 Boyne (HYMOD) 0.79 0.76 0.80 0.83 28.17 29.35 6 1971-1990/1991- Boyne(NAM) 0.76 0.74 0.77 0.78 23.82 25.10 2000 7 Boyne (TANK) 0.70 0.73 0.67 0.75 25.60 27.13 8 Boyne (TOP) 0.69 0.68 0.52 0.57 23.26 24.74 Ensemble Med 0.80 0.78 0.90 0.92 31.78 33.40 9 Suck (HYMOD) 0.78 0.68 0.70 0.68 17.27 18.75 10 1971-1990/1991- Suck (NAM) 0.72 0.63 0.56 0.51 14.68 15.85 11 2000 Suck (TANK) 0.70 0.65 0.61 0.59 17.08 18.45 12 Suck (TOP) 0.68 0.60 0.34 0.31 12.65 14.06 Ensemble Med 0.79 0.69 0.74 0.70 19.24 20.92 13 Blackwater (HYMOD) 0.64 0.74 0.50 0.58 25.18 25.67 14 1971-1990/1991- Blackwater (NAM) 0.63 0.72 0.31 0.40 15.62 16.13 15 2000 Blackwater (TANK) 0.67 0.75 0.59 0.68 33.35 34.09 16 Blackwater (TOP) 0.64 0.71 0.33 0.31 21.77 22.69 Ensemble Med 0.66 0.74 0.68 0.76 36.52 37.32 PI (Width of 90 Prediction interval); CE (No of points with in PI/No of points)

GLUE/BMA BMA GLUE

B Quantify uncertainties that cascade through the climate change impact assessment

Uncertainty Envelope: Experiment design GCM: Weighted based on CCCma CSIRO HadCM3 Climate prediction index Scenarios: Equally likely B2 B2 A2 A2 B2 A2 HyMod HyMod HyMod HyMod HyMod NAM Tank NAM Tank NAM Tank TOP TOP TOP Tank Tank w1, σ 1; w2, σ 2; w3, σ 3; NAM NAM TOP TOP Model: Equally Likely HyMod w4, σ 4 (The weight NAM Tank TOP parameters are revised Simulators: Weighted based on GCM weight) based on Likelihood value Total Hydro Scenario BMA GCM Hydro: Hydrological model uncertainty (parameter & model selection) Scenario: Hydrological + Scenario (A2 & B2) GCM: Hydrological + Scenario (A2 & B2) Total: Uncertainty envelop (Hydrological + Scenario (A2 & B2)+GCM)

Hydrological model uncertainty Moy Boyne Blackwater suck 100 Widh of prediction interval (% of 90 80 70 60 50 40 average flow) 30 20 NAM TANK HYMOD TOP 10 0 Uncertainty due to parameterization and Model selection CCCMA (A2) CCCMA (B2) CISRO (A2) CISRO (B2) HADCM3 (A2) HADCM3 (B2) CCCMA (A2) CCCMA (B2) CISRO (A2) CISRO (B2) HADCM3 (A2) HADCM3 (B2) CCCMA (A2) CCCMA (B2) CISRO (A2) CISRO (B2) HADCM3 (A2) HADCM3 (B2) CCCMA (A2) CCCMA (B2) CISRO (A2) CISRO (B2) HADCM3 (A2) HADCM3 (B2) Prediction interval (% of average 120 Moy Boyne Blackwater Suck 100 80 flow) 60 40 Climate scenarios a)2050-2059 20 0 2020s 2050s 2070s 2020s 2050s 2070s Parameterization Model selection The average width of the PI from parameterization of CRR models is nearly 50% and nearly increased to 70% when Different CRR models are included