Towards monitoring and prediction of severe droughts by integrating numerical simulation and satellite observation

Yohei Sawada

Ph.D Candidate

River and Environmental Engineering Laboratory,

Department of Civil Engineering, the University of Tokyo

2016/3/2 Asian Water Cycle Symposium 2016

1. Introduction – Two major challenges

[Photos from The Telegraph

http://www.telegraph.co.uk]

1.1. Challenge 1: How can we get the holistic view of droughts?

Natural climate variability

Precipitation deficiency, high temperature etc(

Soil water deficiency

Plant water stress, reduced biomass and yield

Reduced stream flow, inflow to reservoirs,

Groundwater deficiency, ((

Economic, Social, and Environmental impacts

[from National Drought Mitigation Center,

University of Nebraska-Lincoln, USA]

See also [Mishra and Singh, 2010]

Mete

oro

logic

al

Ecolo

gic

al

Hydro

logic

al

� Drought is multi-sector and multi-scale phenomena. Couplings between

hydrology and ecology are important to quantify droughts.

1.2. Challenge 2: Monitoring droughts in data-scarce regions.

obs

obs

Initial Condition

Parameter

Optimization

t

e.g

., S

oil

mois

ture

� Even if the model were perfect, we cannot forecast very well without good initial

conditions and model parameters.

� How can we get the observations to improve our forecast in the ungauged

areas??

2. Challenge (1): How can we get the holistic view of droughts?

Natural climate variability

Precipitation deficiency, high temperature etc(

Soil water deficiency

Plant water stress, reduced biomass and yield

Reduced stream flow, inflow to reservoirs,

Groundwater deficiency, ((

Economic, Social, and Environmental impacts

Mete

oro

logic

al

Ecolo

gic

al

Hydro

logic

al

2.1. Ecohydrological model: WEB-DHM-Veg

GBHM(river model)

Coupling

＋

Hydro-SiB(Land surface model)

Dynamic Vegetation Model (DVM)

� WEB-DHM-Veg can simulate soil

moisture, groundwater, river discharge,

and vegetation growth (and their

interactions).

2.2. Strategy of ecohydrological drought quantification

Ecohydrological Model

In-situ observed rainfall JRA25 reanalysis [Onogi et al., 2007]

LAI

Soil Moisture Groundwater

River dischargeSatellite LAI

(AVHRR)

Calibration

& ValidationIn-situ river

discharge

Calibration

& Validation

Agricultural

Drought Index

Nationwide crop production

&

Reports about past droughts

Validation

Hydrological

Drought Index

Drought

Analysis

Drought Indices - Standardized Anomaly Index (SA index) –[Jaranilla-Sanchez et al., 2011]

2.3. Model validation

Drought indices (SA index)

Green:simulated annual peak LAI and Orange:nationwide crop production

� The drought index calculated from the model-estimated annual peak of leaf area

index correlates well with the drought index from nationwide annual crop production.

R =0.89

Dro

ug

ht

Nash = 0.66

R = 0.80

River Discharge at Jendouba site

Blue: Simulated river discharge

Red: Observed river discharge

Bar: Rainfall

[Sawada et al., 2014, Water Resour. Res.]

2.4. Ecohydrological drought analysis on 1988-1989 drought

Drought indices

Blue: River discharge Gray: Groundwater level Green: Leaf Area Index

Dro

ug

ht

Agricultural Drought

Hydrological Drought

� Historic agricultural droughts predominantly occurred prior to hydrological

droughts and the hydrological drought lasted much longer, even after crop

production has recovered.

[Sawada et al., 2014, Water Resour. Res.]

3. Challenge (2): Monitoring droughts in data-scarce regions

obs

obs

Initial Condition

Parameter

Optimization

t

e.g

., S

oil

mois

ture

� How can we get the observations to train the numerical simulation

in the ungauged areas ?

Satellite!!

3.1. Application of Passive Microwave Remote Sensing

Radiative Transfer in microwave region

Radiation from soil � depends on Surface Soil Moisture

Attenuation by canopy

Radiation from canopy

� depend on

Vegetation water content

• Microwave brightness temperature is influenced by surface soil moisture,

vegetation water content, and temperature [e.g., Paloscia and Pampaloni, 1988]

• It is not strongly influenced by atmospheric condition

� By assimilating this data, we can improve the skill of eco-hydrological model to

simultaneously calculate soil moisture and vegetation dynamics.

AMSR-E AMSR2

3.2. Coupled Land and Vegetation Data Assimilation System(CLVDAS)

Ecohydrological

model

Soil moisture

Vegetation(LAI)

Temperature

Radiative

Transfer ModelEstimated TB

Core-Model

Pass1:

Parameter

Optimization

Parameter

Core-Model

Estimated TB

Satellite

observed TB

Schuffled

Complex

Evolution

COST

Pass2:

Data Assimilation

~1year

Soil Moisture, LAI

ensemble

Core-Model

Estimated TB

~5days

Satellite

observed TB

COST

Genetic

Particle

Filter

3.3. Application: Horn of Africa drought

[FAO, 2011]

[Anderson et al., 2012]

� We cannot have the access to many ground observations to develop the

drought prediction system.

3.4. Strategy of ecohydrological drought forecast

2003 2004 2005 2006 2007 2008 2009 2010 2011Observed

meteorological

forcings (satellite-

based)

Microwave land

observation

CLVDAS

(Reanalysis)

Drought

CLVDAS

(Ensemble Stream

Prediction)

CLVDAS

(Real Predicion)

NOAA GFDL

Meteorological

forecast

No in-situ data is used.

3.5. Results (1) 2010-2011 drought in reanalysis

LAI anomaly of 2010-2011 droughts in “reanalysis”.

3.6. Results (3) Predictions: starting from 1 Sep 2010Gray: Climatorogy

Green: Horn of Africa drought (reanalysis)

Leaf Area Index timeseriesC

LV

DA

S

(Ensem

ble

Str

eam

Pre

dic

tion)

CLV

DA

S

(Real P

redic

ion)

[Sawada and Koike, JGR-A, submitted]

� Ecosystem damage of the Horn of Africa drought is predictable 10 months before.

3.6. Results (3) Predictions: starting from 1 Oct 2010Leaf Area Index timeseries

CLV

DA

S

(Ensem

ble

Str

eam

Pre

dic

tion)

CLV

DA

S

(Real P

redic

ion)

Gray: Climatorogy

Green: Horn of Africa drought (reanalysis)

[Sawada and Koike, JGR-A, submitted]

3.6. Results (3) Predictions: starting from 1 Jan 2011Leaf Area Index timeseries

CLV

DA

S

(Ensem

ble

Str

eam

Pre

dic

tion)

CLV

DA

S

(Real P

redic

ion)

Gray: Climatorogy

Green: Horn of Africa drought (reanalysis)

[Sawada and Koike, JGR-A, submitted]

3.6. Results (3) Predictions: starting from 1 Mar 2011Leaf Area Index timeseries

CLV

DA

S

(Ensem

ble

Str

eam

Pre

dic

tion)

CLV

DA

S

(Real P

redic

ion)

Gray: Climatorogy

Green: Horn of Africa drought (reanalysis)

[Sawada and Koike, JGR-A, submitted]

� Ensemble stream prediction (with no meteorological prediction skill) can predict

ecosystem damages to some extent in the short lead time predictions.

4. Towards real-time drought early warning system

Met. Reanalysis

Satellite (AMSR-E, AMSR2, MODIS, GRACE,.....)

CLV

DA

Spast present future

L-V reanalysis

Met. Hindcast Met. Seasonal Forecast

L-V hindcast

L-V real-time forecast

Data

Analy

sis

Tool

river discharge (GRDC)

Soil moisture network

(SCAN, CEOP,5)

Data

base (

DIA

S)

Crop production (FAOSTAT)

Other Data

Ecohydrological Drought analysis[Jaranilla-Sanchez et al., 2011, WRR]

[Sawada et al. 2014, WRR]

Hydrological Drought Index

Agricultural Drought Index

Socio-Economical Drought analysis[Suzuki, 2015]

[Yokomatsu et al., 2015]

SocioEconomical Drought Index

Decision Makers

5. Conclusions

• By explicitly simulating ecosystem damages in addition to

hydrological deficits, we can get the holistic view of

severe drought progress.

• We make it possible to monitor and predict droughts in

the data scarce regions by using the globally applicable

satellite data and data assimilation technology.

� Towards early-warning system of mega-droughts in the data scarce

regions by integrating numerical simulation and satellite observations.