S. Uhlenbrook, J.W. Foppen, I. Masih, S. Maskey, C. Orup, A. - - PowerPoint PPT Presentation

Predicting the Impact of Change –

The need for a better hydrological process understanding through innovative experimental and modeling approaches

• S. Uhlenbrook, J.W. Foppen, I. Masih, S. Maskey, C. Orup,
• A. Saraiva, V. Smakthin, J. Wenninger

20-23 September 2010, HydroPredict’ 2010, 2nd Intern. Interdisciplinary Conference on Predictions for Hydrology, Ecology, and Water Resources Management: Changes and Hazards caused by Direct Human Interventions and Climate Change; Prague, Czech Republic

(1) Intro the world is changing (‘stationary is dead’) (2) Case study ONE: Improving hydrological predictions in the semi-arid Karkheh basin, Iran (3) Case study TWO: DNA – New multi-tracing opportunities to study hydrological flow pathways (4) Case study THREE: The use of stable isotopes to improve our understanding of evaporation fluxes

Ou Outlin line

2 4 3 5 6 1 Global Temperature (°C)

IPCC Projections 2100 AD

N.H. Temperature (°C) 0.5 1

• 0.5

1000 1200 1400 1600 1800 2000

Lower Risk for Instabilities High Risk for Instabilities

It is Getting Warmer!

(NEAA, 2009)

The context – ONE

(NEAA, 2009)

Climate Sensitivity – Best estimate +3C for

2x CO2 pre-industrial, but it can be much higher …

Change in annual runoff by 2041-60 (SRES A1B) – Ensemble of 12 climate models

Source: Kundzewicz et al. (2007); chapter in IPCC (2007)

'climate colonialism'

• A massive land-grabbing

scramble in Africa as foreign companies - some with foreign aid money support - rapidly establish enormous monoculture fields in tropical countries.

Prof Seif Madoffe, SUA Sugar Cane – Kilombera Basin, Tanzania

The context – TWO

Picture from Fairless, 2007, Nature

EI EI ET ET ES ES SS SS QR QR QR QR QS QS P P P P

Impact of land use change on hydrological processes

Short-term dynamics (e.g. interception, flood generation) vs. long-term dynamics (e.g. groundwater recharge, base flow)

     

P Q Q E E Q E E

g g f u T u s s s I I

dt dS dt dS dt dS dt dS

                

 

I I

E

dt dS 

Interception processes

 

s s s

Q E

dt dS

 

Surface water processes       

g g

Q dt dS Groundwater processes

Water Balance Equation: Where:

Root zone moisture processes

 

f u T u

Q E E

dt dS

  

Possible changes in all variables due to climate and/or land changes!!

Global Changes

• Climate (temperature, precipitation, radiation …)
• Land use, land cover

 De-forestation / re-forestation  Urbanisation  Etc.

• Population (amount, density, structure, …)
• Hydraulic works
• Technological development
• Globalisation
• Water use in space and time
• Economic development
• Change of diet (more meat => more water)
• N- and P-fluxes to water bodies
• Pollution (new substances etc.)
• Change in composition of species
• etc. etc. etc.

…. and many interdependencies/feedbacks!

Why is it so difficult to predict hydrological effects of change?

1. Many global changes occur simultaneously with positive or negative (unknow) feedbacks 2. Spatial and temporal scales for hydrological processes are different from scales dominant in other disciplines 3. Hydrological processes are often non-linear or depend on thresholds/tipping points 4. Hydrological extremes (e.g. floods and droughts) do not

• ccur often and are difficult to measure, consequently, good

data sets are usually not available 5. Boundary conditions during hydrological modelling are not clear (i.e. subsurface flows) 6. Hydrological observation methods are insufficient to study hydrological process dynamics (e.g. subsurface flow processes, extreme events etc.)

(1) Intro the world is changing (‘stationary is dead’) (2) Case study ONE: Improving hydrological predictions in the semi-arid Karkheh basin, Iran (3) Case study TWO: DNA – New multi-tracing opportunities to study hydrological flow pathways (4) Case study THREE: The use of stable isotopes to improve our understanding of evaporation fluxes

Ou Outlin line

The Karkheh basin, Iran

Some basic facts and figures



Drainage area: 50,764 km2



More than 80 % is mountainous



Divided into five sub-basins



Mediterranean climate: Cool and wet winter; dry and hot summers



Precipitation 450 mm/year, range: 150 mm to 750 mm

Water allocations in 2001 (4949 MCM) Irrigation, 4149 Environment, 500 Others, 14 Domestic, 262 Industry, 23

Water allocations in 2025 (8903 MCM) Irrigation, 7416 Environment, 500 Others, 512 Industry, 113 Domestic, 362

Source: JAMAB 2006

Improving precipitation input in rainfall- runoff modeling using SWAT

• The current way of climatic data input in SWAT is

rather simple

One station nearest to the centroid of a catchment

Gauge nearest to the centroid may not be the best representative This can undermine the full use of available data (e.g. if two stations in a sub-catchment, only one will be used)

• Quality of the climatic data input will has serious

implications for the model parameterization and quality of (spatial and temporal) the results

Preparation of areal precipitation input

1) Rain gauge data 2) Gauge location 3) DEM /Elevation 4) Sub-basin ID Interpolation using IDW including elevation weighting Cross validation Areal average for sub-catchment Virtual rain gauge data Input for each sub-catchment

Masih et al., JAWRA; in review

Comparison of the input precipitation:

Case II (areal precipitation) vs. Case I (station data): Spatial view

High spatial variability, mainly influenced by topography (left) The precipitation difference in Case II compared to Case I ranged from -40 to 40 % (right)

Sub-catchment precipitation (Case II) Precipitation difference (Case II vs. Case 1)

Comparison Case II vs. Case I: Temporal view

Divergent variations by sub-catchment, illustrated by four selected cases. Precipitation dynamics in Case II could be different in many respects. Daily values can be higher/lower. They also show clear pattern in extreme values: 1) lower P events can be totally missed

• ut be a single rain gauge;

2) extremes in Case II are comparatively small in most cases, though could be other way around for some sub-catchments and P events.

Sub-catchment ID: 1 20 40 60 80 100 20 40 60 80 100 P Case II (mm/day) P Case I (mm/day) Sub-catchment ID: 1 50 100 150 200 250 300 50 100 150 200 250 300 P Case II (mm/month) P Case I (mm/month) Sub-catchment ID: 1 200 400 600 800 1000 200 400 600 800 1000 P Case II (mm/year) P Case I (mm/year) Sub-catchment ID: 63 20 40 60 80 100 120 20 40 60 80 100 120 P Case II (mm/day) P Case I (mm/day) Sub-catchment ID: 63 50 100 150 200 250 300 350 50 100 150 200 250 300 350 P Case II (mm/month) P Case I (mm/month) Sub-catchment ID: 63 400 800 1200 1600 400 800 1200 1600 P Case II (mm/year) P Case I (mm/year) Sub-catchment ID: 33 20 40 60 80 100 20 40 60 80 100 P Case II (mm/day) P Case I (mm/day) Sub-catchment ID: 33 50 100 150 200 250 300 50 100 150 200 250 300 P Case II (mm/month) P Case I (mm/month) Sub-catchment ID: 33 200 400 600 800 1000 200 400 600 800 1000 P Case II (mm/year) P Case I (mm/year) Sub-catchment ID: 43 20 40 60 80 100 20 40 60 80 100 P Case II (mm/day) P Case I (mm/day) Sub-catchment ID: 43 50 100 150 200 250 300 50 100 150 200 250 300 P Case II (mm/month) P Case I (mm/month) Sub-catchment ID: 43 200 400 600 800 1000 200 400 600 800 1000 P Case II (mm/year) P Case I (mm/year)

Daily Monthly Annual Masih et al., JAWRA; in review

SWAT calibration and performance evaluation



Rigorous calibration approach using both manual and automatic procedure

(SUFI-2, Abbaspour et al., 2007)



Daily climatic data of 1987-2001 (Precipitation: 41 stations; Temperature: 11 stations)



Performance evaluation: NSE, R2 and annual volume balance



15 stream flow gauges across the Karkheh River System



Temporally at daily, monthly and annual time scales, over period of 1987-2001

(Calibration: Oct 1987-Sep1994; Validation: Oct 1994-Sep 2001)

Comparison of stream flow simulations under Case I & II

Comparison (Calibration: Oct 1987-Sep 1994)

• 1.1
• 0.8
• 0.5
• 0.2

0.1 0.4 0.7 1.0 42620 39940 20863 10860 9140 5370 2320 1590 1460 1260 1130 844 800 776 590

Drainage area (km2) Daily NSE (-)

Case I (Station data) Case II (Interpolated data)

Masih et al., JAWRA; in review

Comparison of streamflow simulations under Case I & II

Comparison (Validation: Oct 1994-Sep 2001)

• 2.9
• 2.6
• 2.3
• 2.0
• 1.7
• 1.4
• 1.1
• 0.8
• 0.5
• 0.2

0.1 0.4 0.7 1.0 42620 39940 20863 10860 9140 5370 2320 1590 1460 1260 1130 844 800 776 590

Drainage area (km2) Daily NSE (-)

Case I (Station data) Case II (Interpolated data)

Masih et al., JAWRA; in review

Uncertainty analysis using SWAT-CUP: Summary of results

P-Factor indicates the percentage of observed data bracketed by 95PPU band. The achieved values are in reasonably good range (e.g. >0.5 in most cases)

P-Factor based on daily calibration (1988-94) 0.0 0.2 0.4 0.6 0.8 1.0 Doabe Merek Khers Abad Aran Firoz Abad Pole Chehre Ghore Baghestan NoorAbad Holilan Sarab Seidali Kaka Raza Cham Injeer Afarineh Chalhool Pole Dokhtar Jelogir PayePole P-Factor (-)

Uncertainty analysis using SWAT-CUP: Summary of results

R-Factor indicates the width of the 95PPU band. The achieved values are in reasonably good range (<0.5 in most cases)

R-Factor based on daily calibration (1988-94) 0.0 0.2 0.4 0.6 0.8 1.0 Doabe Merek Khers Abad Aran Firoz Abad Pole Chehre Ghore Baghestan NoorAbad Holilan Sarab Seidali Kaka Raza Cham Injeer Afarineh Chalhool Pole Dokhtar Jelogir PayePole P-Factor (-)

Uncertainty analysis using SWAT-CUP: Example results

Jelogir at the Karkheh River (1988-1994)

300 600 900 1200 1500 Jan-88 Jul-88 Jan-89 Jul-89 Jan-90 Jul-90 Jan-91 Jul-91 Jan-92 Jul-92 Jan-93 Jul-93 Jan-94 Jul-94

Month Discharge (m3/s)

M95PPU Observed

Most of the observed data fall well within 95PPU band.

Uncertainty analysis using SWAT-CUP: Example results

Most of the observed data fall well within 95PPU band.

Jelogir at the Karkheh River (1988-1994) 100 200 300 400 500 600 700 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep

Month Discharge (m3/s)

M95PPU Observed

Main findings from this case study

• Areal precipitation, based on interpolation of the available

station data, improved SWAT model

• Results were strongly influenced by the spatial extent and

the station density/spatial distribution of the rain gauges

 Smaller catchments (600-1600 km2) showed noteworthy improvements  Larger catchments (>5000 km2) showed comparable performance

• Uncertainty analysis applying SUFI-2 algorithm was used

(Abbaspour et al., 2007)

• Next steps:

 Testing of other (semi-)distributed models  Use of other input data, e.g. interpolation methods, radar data and satellite observations  More attention to model parameterization and uncertainty analysis  Evaluating the downstream impacts of increasing water consumption in the upstream rain-fed area

(1) Intro the world is changing (‘stationary is dead’) (2) Case study ONE: Improving hydrological predictions in the semi-arid Karkheh basin, Iran (3) Case study TWO: DNA – New multi-tracing opportunities to study hydrological flow pathways (4) Case study THREE: The use of stable isotopes to improve our understanding of evaporation fluxes

Ou Outlin line

Background

Synthetic DNA in multi-tracing hydrological processes

• DNA is a nucleic acid  contains genetic

instructions

• DNA has got unique inherent coding abilities
• Multiple DNA can be designed and produced in the

laboratory

• Each DNA can be determined specifically in solution

using quantitative polymerase chain reaction (qPCR)

• First experiments in groundwater studies (Sabir et

al., 1999) – were only qualitative and showed how DNA can be used as a marker

• Two case studies were carried out in surface water

and laboratory column experiment between May and July 2010

100 m downstream

1) Merkske stream, The Netherlands

600 m downstream

• Discharge of 50 l/s
• 6 kg of NaCl injected
• 6 DNA (each 1ml of 1.67 μ M) injected
• All DNA traced downstream
• Similar BTC as that of NaCl

Results – Surface water

Elapsed Time (Hrs)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

DNA concentration (pM)

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

NaCl concentration (mg/l)

50 100 150 200 250 DNA2 DNA3 DNA4 DNA5 DNA6 DNA8 NaCl

Elapsed Time (Hrs)

1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

DNA concentration (pM)

0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016

NaCl concentration (mg/l)

10 20 30 40 50 60 70 DNA2 DNA3 DNA4 DNA5 DNA6 DNA8 NaCl

Foppen et al., in prep.

PERCENTAGE RECOVERY AT LOCATION 2 – 600m DNA2 DNA3 DNA4 DNA5 DNA6 DNA8 25% 64% 73% 82% 85% 59%

Elapsed time (Hr)

1 2 3 4 5 6 7

NaCl concentration (mg/l)

100 200 300 400

DNA concentration (pM)

0.00 0.02 0.04 0.06 0.08 NaCl conc. at Loc. 1 (115m downstream) NaCl conc. at Loc. 2 (620m downstream) NaCl conc. at Loc. 3 (1200m downstream) DNA conc. at Loc. 1 (115m downstream) DNA conc. at Loc. 2 (620m downstream) DNA conc. at Loc. 3 (1200m downstream)

Results – Surface water

2) Strijsbeekse beek stream, The Netherlands

• Discharge of 40 l/s
• 6 kg of NaCl injected
• 6 DNA (each 1ml of 1.67 μ M)
• All DNA traced downstream
• Similar BTC as that of NaCl

115m 620m 1200m

Foppen et al., in prep.

Results – Transport in porous media

BTC of NaCl and DNA , laboratory column – pure quartz sand

• NaCl influent concentration - 0.5 g/l
• DNA influent concentration - 0.01 μ M
• 4 PV of NaCl and DNA injected at

pore water velocity of 0.4 cm/min

• Similar BTC as that of NaCl  DNA

travels with water

• Kinetic attachment, and not retarded

– reduced peak concentration

• Slow detachment – seen in recession

limb

Pore Volume (-)

5 10 15 20

C/C0

1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3

E/E0

0.001 0.01 0.1 1 DNA2 DNA3 DNA4 DNA5 DNA6 DNA8 NaCl

Foppen et al., in prep.

Main findings - DNA as Tracer

• DNA travels with water and can be detected at

very low concentrations consisting of multiple DNAs

• Very small quantities were required as input
• Suitable as tracers for multi-tracing experiments
• More experiments (lab and field) needed to

further understand its transport properties

(1) Intro the world is changing (‘stationary is dead’) (2) Case study ONE: Improving hydrological predictions in the semi-arid Karkheh basin, Iran (3) Case study TWO: DNA – New multi-tracing opportunities to study hydrological flow pathways (4) Case study THREE: The use of stable isotopes to improve our understanding of evaporation fluxes

Ou Outlin line

Transpiration Evaporation Soil moisture sensors Data Logger Soil Balance Percolation Rhizon water samplers

Better Understanding of Evaporation Fluxes using Environmental Isotopes

Wenninger et al., 2010; PCE

• Evaporation is the driving factor in isotopic fractionation
• Transpiration and percolation do not cause fractionation
• Quantification between fractionating and non-fractionating losses
• Conservation of mass and isotopes

Isotope Mass Balance

p; precipitation t; transpiration v; evaporation z; percolation i; f f; final WC i; initial WC

with:

(e.g. Robertson et al. 2006, J. of Hydrol.)

Lysimeter 2 36.00 37.00 38.00 39.00 40.00 41.00 42.00 02/02/2009 07/02/2009 12/02/2009 17/02/2009 22/02/2009 27/02/2009 Date Weight [kg] 500 1000 1500 2000 2500 3000 Irrigation, Percolation [ml] irrigated water (ml) percolated water (ml) corrected Weight (kg)

Measured water fluxes: Lysimeter A (+vegetation cover)

Results Lysimeter Experiments

Irrigation (mm) Percolation (mm) 116 13

Wenninger et al., 2010; PCE

Lysimeter 2 2 4 6 8 10 12 14 16 18 20 02/02/2009 07/02/2009 12/02/2009 17/02/2009 22/02/2009 27/02/2009 Date Evapotranspiration, Percolation [mm] 10 20 30 40 50 Irrigation [mm] Irrigation Percolation Evapotranspiration

Measured water fluxes: Lysimeter A (+vegetation cover)

Irrigation (mm) Percolation (mm) Evapotranspiration (mm) 116 13 134

Results Lysimeter Experiments

Wenninger et al., 2010; PCE

Isotope Depths Profiles and Evaporation Line

Wenninger et al., 2010; PCE

• The irrigation water is in the range of the GMWL
• Soil water samples are isotopically heavier and move along the

evaporation lines

Meteoric Water Line and Evaporation Line

y = 3.47x - 20.9 R2 = 0.97 y = 3.75x - 18.7 R2 = 0.98

• 54
• 34
• 14

6 26

• 8
• 6
• 4
• 2

2 δ2H δ 18O

2 )

GMWL

Lysimeter A: slope = 3.75 Lysimeter B: slope = 3.47 δ δ Lysimeter A:

Irrigation Soilwater Percolation δ δ

2

Lysimeter B:

Irrigation Soilwater Percolation

Wenninger et al., 2010; PCE

Comparison between different evaporation estimations: (a) measured using hydrometric data and HYDRUS 1D, and (b) calculated using isotope mass balance.

(a) (b) (a) (b)

New way to estimate evaporation fluxes?!

Lysimeter imeter A Lysimeter imeter B

Evaporation E (mm) 19 77 Transpiration T (mm) 99 T/Etotal ratio 84%

 Area: 46,800 km2  Semi-arid climate

• Rainfall ~ 740 mm/a
• Epot ~ 1900 mm/a

Mhlume Estates

 Irrigated sugar cane

Water scarcity

INTRODUCTION – STUDY AREA

(Adapted from Carmo Vaz et al., 2003)

Results - Climate data

5 10 15 20 25 30 35 10 20 30 40 50 60 31/10/09 07/11/09 14/11/09 21/11/09 28/11/09 05/12/09 12/12/09 19/12/09 26/12/09 02/01/10 09/01/10 16/01/10 23/01/10 30/01/10 Radiation (MJ/m2) Precip itation & Et (mm) Rain (mm) Et (mm) Radiation (MJ/m2)

Results - Soil Moisture

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 26/11/09 06/12/09 16/12/09 26/12/09 05/01/10 15/01/10 25/01/10

Drip Field - M422.01 - Trench 1

Avg Port 1 Avg Port 2 Avg Port 3 AvgPort 4 AvgPort 5 MaxPort 1 MaxPort 2 MaxPort 3 MaxPort 4 MaxPort 5 Rain (mm) Net Irrig. M422(mm)

Results - Soil Moisture

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 26/11/09 06/12/09 16/12/09 26/12/09 05/01/10 15/01/10 25/01/10

Furrow Irrigation field - M430.01 - Trench 2

Avg Port 1 Avg Port 2 Avg Port 3 AvgPort 4 MaxPort 1 MaxPort 2 MaxPort 3 MaxPort 4 Rain (mm) Net Irrig. M430 (mm)

Results - Sap Flow / Transpiration

100 200 300 400 500 600 700 800 900 1000 30/12 31/12 01/01 02/01 03/01

Sap Flow

Flow_1 g/hr Flow_2 g/hr Flow_3 g/hr Flow_4 g/hr Average F1,F2,F3

Correlation Sapflow vs ET Plant density ~ 130 000/ ha LAI ~ 4 to 7

Concluding Remarks

 The world is changing – Hydrology too (‘stationary is dead’)  Global changes (incl. CC) are impacting the hydrological cycle; i.e. often ‘acceleration of the water cycle’, but not consistent world-wide  SWAT application in Karkheh basin: Need for new model? Innovate existing ones?  New experimental methods are needed!

• DNA offers new possibilities to trace flow pathways
• Potential of environmental isotopes to measure evaporation fluxes

demonstrated through lab experiments

• First interesting field results from Swaziland – more to come …

Progress in science depends on new techniques, new discoveries and new ideas, probably in that order (S. Brenner, 1980)