hydrology.asu.edu with different hydrologic models · increase in the fut period, with the highest...

SWAT tRIBS WASIM−20

−15

−10

−5

0

Change in mean annual SWC (%)

ECH−RCA

ECH−REM

ECH−RMO

HCH−RCA

SWAT tRIBS WASIM0

5

10

15

20

25

30

Mean annual SWC (%)

REF

FUT

Climate change effects in a Mediterranean basin

with different hydrologic models

WaSiM-ETH, developed by Schulla and Jasper, is a process based and fully distributed hydrological model.

The applied version uses the Richards equation for the water flow within the unsaturated soil zone.

The model is grid based and has a modular structure to represent the physical processes, as described in Fig. 4.

###

##

#

# #

#

#

# #

")

´

km30

Outlet section

..

.

Elevation

(m a.s.l.)1824

0

G Centroid of 25-km grid.

! Centroid of 5-km grid.

# Rain gauges.

Sub-basins

(a) (c)

(b)##

#

#

#

#

#

#

#

#

#

#

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

!

G

G

")

497500

497500

507500

507500

517500

517500

527500

527500

43

57

50

0

43

57

50

0

43

67

50

0

43

67

50

0

43

77

50

0

43

77

50

0

43

87

50

0

43

87

50

0

´

5km

Elevation (m a.s.l.)

963

66

stream

network

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

G

")

1

2

Monica Piras(1), Giuseppe Mascaro(2), Roberto Deidda(1) , Enrique R. Vivoni(3), Pierluigi Cau(4), Pier Andrea Marras(4), Swen Meyer(5), and Ralf Ludwig(5)

Three hydrologic models (HMs) with different characteristics and schematizations are used to simulate the hydrological response due to climate change scenarios in RMB: SWAT, tRIBS and WASIM.

The 3 HMs accordingly predict a decrease in mean annual

discharge Q (Fig. 7a and b). Variation depends also on CMs.

Fig. 5 - (a)-(b) Mean annual MAP (a) and T (b) in REF and FUT periods.

(c)-(d) Relative change in monthly MAP (c) and T (d).

Two time slices: Reference (REF 1971-2000) and Future

(FUT 2041-2070) periods.

All CMs predict decreasing mean annual MAP (mean areal P), and

increasing mean annual temperature T (Fig. 5 a,b) in FUT period.

MAP is predicted to slightly increase in winter months (Fig. 5c) and

decrease in the other seasons, while T is expected to increase

throughout the year (Fig. 5d).

HCH-RCA presents the largest variations for both P and T.

High differences are found in the ETP estimations due to the

different computational schemes. ETP is predicted to slightly

increase in the FUT period, with the highest rise predicted by

WaSiM model (Fig. 6). HCH-RCA presents the largest variations.

Fig. 6 - (a) Mean annual ETP in REF and FUT periods used in the 3 HMs.

(b) Relative change in annual ETP specific for each CM.

The real evapotranspiration ETR diminuishes in FUT according to

the 3 HMs (Fig. 7c and d). Variation depends on CMs .

Deidda et al. (2013) audited 14 combinations of general (GCM) and

regional (RCM) climate models, based on the A1B emission scenario

from the ENSEMBLES project (http://ensembles-eu.metoffice.com),

for this study.

The four best performing

combinations

of two GCMs and

three RCMs were selected,

indicated as CMs.

Climatological center and model Acronym

GCMs

Hadley Centre for Climate Prediction, Met Office, UK

HadCM3 Model (high sensitivity) HadCM3 HCH

Max Planck Institute for Meteorology, Germany ECHAM5 / MPI OM

ECHAM5 ECH

RCMs

Swedish Meteorological and Hydrological Institute (SMHI), Sweden RCA Model

RCA RCA

Max Planck Institute for Meteorology, Hamburg, Germany

REMO Model REMO REM

Koninklijk Nederlands Meteorologisch Instituut (KNMI), Netherlands RACMO2 Model

RACMO RMO

1

ECHAM5

RCA

REMO

RACMO

ECH-RCA

ECH-REM

ECH-RMO

HadCM3 RCA

GCMs RCMs

HCH-RCA

Acronym ofcombinations

tRIBS is a fully distributed, physically-based HM, using Triangulated Irregular Networks to represent topography (Fig. 3).

Fig. 3 - tRIBS application scheme in the RMB (Mascaro et al., 2013): downscaling strategies and generation of hourly forcing for (a) precipitation, P, and (b) potential evapotranspiration, ET

P. (c) Hydrologic processes

represented in tRIBS model over a complex triangulated terrain (Ivanov et al., 2004) and performances of tRIBS calibration and validation in RMB.

Jan 06th Jan 07th Jan 08th0

5

10

15

20

25

30

t (d)

Pre

cip

itatio

n (m

m/d

)

Daily Rainfall

0

0.1

0.2

0.3

0.4

t (h)

ET

0 (m

m/h

)

Apr 28th Apr 29th Apr 30th

h

Hourly Evapotranspiration

0 0.5 1 1.5 2 2.50

0.5

1

1.5

2

3

3.5

4

2.5

β = e -1

Prec. (mm/h) in 104 km x 104 km x 6 h

c

ST

RA

IN m

odel

(a)

0 4 8 12 16 200

0.05

0.1

t (h)

fm(h

)

April

0

1

2

3

4

5

6

t (h)

Pre

cip

itatio

n (m

m/h

)

Jan 06th Jan 07th Jan 08th

Hourly Rainfall

Apr 28th Apr 29th Apr 30th0

0.5

1

1.5

2

2.5

3

t (d)

ET

0 (m

m/d

)d

Daily Evapotranspiration

Dim

ensio

nle

ss

functio

ns

(b)

Time scale Calibration NSC

Min, Mean, Max

Validation NSC

Min, Mean, Max

Daily -3.53, 0.07, 0.61 -0.99, 0.02, 0.42

Weekly -5.50, 0.46, 0.83 -0.72, 0.13, 0.47

Monthly -0.06, 0.55, 0.89 0.30, 0.25, 0.74

(c)

Our study site is the Rio Mannu basin (RMB), located in the

southern Sardinia, Italy (Fig.1) and included within the CLIMB EU

FP7 project (Ludwig et al., 2010):

Fig. 1 - Location of the RMB within (a) Italy and (b) the island of Sardinia. (c) Digital elevation model (DEM) of the RMB including UTM coordinates. Panels (b) and (c) also report the centroids of the original and disaggregated climate model grids and (c) two sub-basins boundaries (violet lines).

Hydrologic Models

TIN-based Real Time Integrated Basin Simulator tRIBS

Study Area

Water Flow and Balance Simulation Model (WaSiM)

Soil and Water Assessment Tool, SWAT

SWAT is a hydrological, semi-distributed model that functions on a continuous time step. Model components include weather, hydrology, channel and pond/reservoir routing. Fig 2 shows its performance when applied on the RMB (Marras et al., 2014).

Hydrologic Impacts

Climate Simulations

Changes in climate forcing

Monthly variations

Conclusions

References

SWAT tRIBS WASIM0

100

200

300

400

500

Mean annual ETR (mm)

REF

FUT

SWAT tRIBS WASIM−20

−15

−10

−5

0

Change in mean annual ETR (%)

ECH−RCA

ECH−REM

ECH−RMO

HCH−RCA

SWAT tRIBS WASIM

−70

−60

−50

−40

−30

−20

−10

0

Change in mean annual Q (%)

ECH−RCA

ECH−REM

ECH−RMO

HCH−RCA

SWAT tRIBS WASIM0

20

40

60

80

100

120

Mean annual Q (mm)

REF

FUT

DICAARUniversità

di

Cagliari

hydrology.asu.edu

Thursday, June 25

Session G1 - Poster

Mascaro G., Piras M., Deidda R., and Vivoni E. R.: Distributed hydrologic modeling of a sparsely monitored basin in Sardinia, Hydrol. Earth Syst. Sci., 17, 4143–4158, doi:10.5194/hess-17-4143-2013, 2013.

Ludwig R., et al.: Climate-induced changes on the hydrology of Mediterra-nean basins - A research concept to reduce uncertainty and quantify risk, Fresen. Environ. Bull., 19 (10 A), 2379–2384, 2010.

Ivanov V. Y., Vivoni E. R., Bras R. L., and Entekhabi D.: Catchment hydrologic response with a fully-distributed triangulated irregular network model, Water Resour. Res., 40(11), doi:10.1029/2004WR003218, 1–23, 2004.

Deidda R., Marrocu M., Caroletti G., Pusceddu G., Langousis A., Lucarini V., Puliga M., and Speranza A.: Climate model validation and selection for hyd-rological applications in representative Mediterranean catchments, Hydrol. Earth Syst. Sci., , 17, 5041-5059, doi:10.5194/hess-17-2013, 2013.

Fig. 4 - Model structure of the WaSiM-ETH (Schulla, 2015).

Me

an

an

nu

al te

mp

era

ture

(°C

)

ECH-RCA ECH-REM ECH-RMO HCH-RCA0

5

10

15

20

25

REF

FUT

+1.9°C +1.9°C +1.9°C+3.0°C

Me

an

an

nu

al M

AP

(m

m)

ECH-RCA ECH-REM ECH-RMO HCH-RCA0

100

200

300

400

500

600

700

800

REF

FUT

-12% -7% -10% -21%

Ch

an

ge

in

mo

nth

ly M

AP

(%

)

Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug−80

−60

−40

−20

0

20

40

ECH-RCAECH-REMECH-RMOHCH-RCA

Ch

an

ge

in

mo

nth

ly T

(%

)

0

5

10

15

20

25

30

35

40

ECH-RCA

ECH-REM

ECH-RMO

HCH-RCA

Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug

(b)(a)

(d)(c)

Also the soil water content SWC is predicted to diminuish in FUT

(Fig. 7e and f ).

Fig. 7 - (a) Mean annual Q in REF and FUT periods simulated by the 3 HMs.

(b) Relative change in annual Q specific for each CM. (c) and (d) Same as (a)

and (b) for ETR. (e) and (f) Same as (a) and (b) for SWC.

SWAT tRIBS WASIM0

5

10

15

20

25

Change in mean annual ETP (%)

ECH−RCA

ECH−REM

ECH−RMO

HCH−RCA

(b)

SWAT tRIBS WASIM0

500

1000

1500

Mean annual ETP (mm)

REF

FUT(a)

The highest variations in Q, ETR and SWC are found for HCH-RCA.

The 3 HMs present larger variations considering the seasonal

scale (Fig. 8) as compared to the annual scale (Fig. 7).

Q is predicted to decrease in FUT in all months except for January

and February (Fig. 8b).

(b)(a)

(d)(c)

(f)(e)

Fig. 8 - (a) Mean monthly Q in REF (solid lines) and FUT (dashed lines) periods

simulated by the 3 HMs. (b) Relative change in mean monthly Q for each HM.

(c) and (d) Same as (a) and (b) for ETR. (e) and (f) Same as (a) and (b) for SWC.

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−80

−60

−40

−20

0

20

Change in mean monthly Q (%)

SWAT

tRIBS

WASIM(b)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug0

5

10

15

20

25

Mean monthly Q (mm)

SWAT

tRIBS

WASIM

(a)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−30

−20

−10

0

10

Change in mean monthly ETR (%)

SWAT

tRIBS

WASIM (d)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug0

20

40

60

80

Mean monthly ETR (mm)

SWAT

tRIBS

WASIM

(c)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−40

−30

−20

−10

0

Change in mean monthly SWC (%)

SWAT

tRIBS

WASIM

(f)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug0

10

20

30

40

Mean annual SWC (%)

SWAT

tRIBS

WASIM

(e)

ETR increases only in December and January according to WaSiM.

Largest descrease in summer months predicted by WaSiM (Fig. 8d).

Using 4 CM signals (coherently predicting lower mean annual

precipitation and higher mean temperatures in the future period

2041-2070) as forcing of 3 HMs with different schematizations in

the RMB leads to the following conclusions:

All HMs indicate a negative trend of both mean annual discharge

and mean annual real evapotranspiration (the latter likely due to

drier soil moisture conditions), with minor differences across

the HMs.

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−100

−50

0

50

Change in mean monthly Q (%)

S

M

SWAT

tRIB

WASI

(b)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−2

0

2

4

6

8

Mean monthly Q (mm)

S

SWAT

tRIB

WASIM

(a)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug−80

−60

−40

−20

0

20

Change in mean monthly Q (%)

SWAT

tRIBS

WASIM

(d)

Sep Oct NovDec Jan FebMar AprMay Jun Jul Aug0

10

20

30

40

50

60

Mean monthly Q (mm)

SWAT

tRIBS

WASIM

Sub-basin 2 Sub-basin 2(c)

The mean monthly Q at the outlet section of two sub-basins with

diverse terrain and soil properties (violet boundaries in Fig. 1c)

shows different trends among the HMs (Fig. 9).

Fig. 9 - (a) Mean monthly Q in REF (solid lines) and FUT (dashed lines) periods

simulated by the 3 HMs in sub-basin 1. (b) Relative change in mean monthly

Q in sub-basin 1 for each HM. (c) and (d) Same as (a) and (b) for sub-basin 2.

Sub-basins

Differences in the HMs outputs become more relevant when

looking at higher time resolutions, e.g. at the monthly seasonal

cycles, revealing the uncertainty of the hydrological component in

climate change impact assessments.

The analyses at sub-basin scale displays that, despite the different

process schematization, all HMs were able to reliably represent the

influence of terrain and soil properties in terms of basin response.

Calibration: 1924-1937

Validation: 1943-1963

NS: 0.68

P-factor: 0.7

R-factor: 1.03

NS: 0.60

P-factor: 0.6

R-factor: 0.61

95PPU

Observed

Best estimation

95PPU

Observed

Best estimation

10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 (months)

10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 (months)

12

10

8

6

4

2

0

Q (

m3/s

)

15

12

9

6

3

0

Q (

m3/s

)

Schulla J., Model description WaSiM (Water balance Simulation Model), Hydrology Software Consulting, 2015.

Marras P., Muroni D., Manca S., Soru C., Pusceddu G., Marrocu M., and Cau P.: The SWAT model and a web-based information system to assess the water balance of Sardinia (Italy), 2014 SWAT International Conference, 2014.

Fig. 2 - Results of calibration and validation of the SWAT HM in the RMB.

Sub-basin 1 Sub-basin 1