1.Review of global crop ET methods and estimates 2.MOD16 vs SEBAL results (progress report) 3.Priorities for global crop ET map 4.Planned activities Trent.

1.Review of global crop ET methods and estimates2.MOD16 vs SEBAL results (progress report)3.Priorities for global crop ET map4.Planned activities

Trent BiggsSan Diego State University

Four methods for calculating ET from croplands:

1.Back-calculate from crop production2.Crop coefficient method, crop/hydrological models3. Vegetation-based ET, satellite data (MOD16)4. Temperature-based ET, satellite data (SEBAL)

Global crop water use (total ET)Reference km3/yr Method

Postel, 1996 7400 1. Global water productivity, NPP

Rockstrom et al 1999 6800 1. FAO production and Crop water productivity

Rockstrom and Gordon 2001 6700 1. Biome-specific WUE, FAO production

Chapagain and Hoekstra, 2004 6400 1. Yield and national-average crop water productivity

Liu and Yang, 2010 5900 2. EPIC (Kc), CRU climatology, SAGE crop map

Mekonnen and Hoekstra 2010 2. Kc, soil moisture model, CRU, SAGE crops

Global irrigated crop water use (irrigation demand)Reference km3/yr Method

Doll 2002 1092 2. Kc, CRU climatology

Liu and Yang, 2010 952 2. Kc (EPIC), CRU climatology, SAGE crop map

Mekonnen and Hoekstra 2010 2. Kc, soil moisture model, CRU, SAGE crops

Vegetation-basedMOD16

Energy or temperature-based

SEBAL

Leaf-area index

Net radiation

Land surface temperature

Penman-Monteith equation ET = f(Rn, LAI, wind, temp, rH) 1 degree meteorological grids

Energy balance equation ET = f(Rn, LST) Met station data Automated

How much leaf area is on the surface?

How cool is the surface?

ET estimation from satellite imagery: Two approaches

1.Review of global crop ET estimates2.MOD16 vs SEBAL results (progress report)3.Priorities for global map4.Planned activities

Summary of progress since last WG meeting:

1. Automated SEBAL: Recoded in R from pythonMessina and Biggs in revision, submit to JoH September 2013

2. MOD16 vs SEBAL, LSWI, NDVI: Conference presentationsBiggs, T. W., & Messina, A. (2013). Global ET datasets: Comparison with automated SEBAL in irrigated and rainfed ecosystems. In Association of American Geographers. Los Angeles, CA.

Biggs, T. W., Marshall, M., Melton, F., Lund, C., & Adhikari, D. (2013). INVITED. Global evapotranspiration datasets: An evaluation of vegetation and energy-based methods for irrigated areas. In American Geophysical Union Fall Meeting. San Francisco, CA.

3. MOD16 vs SEBAL vs fluxtowers, CATime series of NDVI, LSWI, NDVI at flux sitesGMAO radiation data downloaded, Rn calculated

4. (new) Interannual variability in irrigated area, KrisnhaBiggs, T.W., Gumma, M., Scott, C. (2013) Are small reservoirs important? The regional sociohydrology and climatic sensitivity of irrigation systems in southern India. In American Geophysical Union Fall Meeting. San Francisco, CA.

Global ET product: MOD16 (Mu et al, 2011) 2000-2010, 8-day

Validation focus:Natural veg, rainfed ag.(95% global ET)

RAINFED VEGETATIONSeasonal averages match Intraseasonal differences at end of post-monsoon season

IRRIGATED VEGETATION: MOD16 ET < SEBAL ET in irrigated areas

When and why does MOD16 underestimate ET in irrigated areas cf SEBAL? MOD16 soil evap = 0 if RH < 70% in GMAOHypothesis: Differences should be largest at beginning of growing season in dry season, when NDVI is low but soil evaporation is high (flooding in rice). Differences should be less at maximum biomass

Photo credit: Halim, N.S. 2010, Agriculture--meeting the water challenge

Large differencebetween MOD16 and SEBAL

Small difference

Test with Land surface wetness index (LSWI) =(NIR-SWIR)/(NIR+SWIR) Chandrasekar et al 2010

Photo credit: P. Thenkabail, 2010, International Water Mgmt Institute

First cropSecond crop

Not flooded, low NDVI , low RHFlooded,

low NDVI

Does the hypothesis hold? Yes— SEBAL > MOD16 when flooded But SEBAL ET is also high late in the cropping season….

After harvest but soil moisture still high?

Sugarcane: Difference largest at end of dry season (April-May)

Not flooded, low NDVI,low RH

How important is accurate estimation of evaporation from open water and wet soil?

Bangladesh field site, eddy flux correlation tower (Hossen et al, 2012) T = 64-70% of ET E = 30-36% of ET

E = 30-40% (Bouman et al, 2005)

CA flux tower sites2011

Green NDVIBlue LSWIBlack MOD16ETGrey SEBALET

Other sites?

Add Kc prediction using MOD16 PET

1.Review of global crop ET estimates2.MOD16 vs SEBAL results (progress report)3.Priorities for global map4.Planned activities

Proposed approach to global cropland ET mapping

1. MOD16 in rainfed croplands (and irrigated corn, wheat, soy?)

2. Kc using MOD16 potential ET in irrigated areas

3. SEBAL to test Kc assumptions by crop type in select semi-arid areas

Key needs: Crop type map (esp rice) Irrig vs non-irrig map (fractional ideal) **Crop calendar**

Heterogeneous pixels (irrig fraction best)

Groundwater irrigation in southern India:

How map irrigated area and ET?

Will global product have irrigated fraction associated with each class?

1.Review of global crop ET estimates2.MOD16 vs SEBAL results (progress report)3.Priorities for global map4.Planned activities

- Comparison of MOD16 and SEBAL at flux towers (with Marshall, NASA)- Publication of automated SEBAL algorithm/results for southern India- Interannual variability in KB, southern India- Proposal preparation…

1950 1960 1970 1980 1990 2000 20100

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000Tanks

Project Canals

Tube Wells

Other Wells

Other Sources

Irrigation by source, Rangareddy District

Irrigated area, ha

Using precip, can determine climate sensitivity of irrigated area

Tanks:

Climate sensitivity of well-irrigated area

Evidence for changes over 1955-2011?

1955-1990: Not sensitive

1990-2011: Erratic, fluctuating

Pe

rce

nt

Ch

an

ge

in

Irri

ga

ted

Are

a

Western Ghats

KrishnaRiverDelta

How can water productivity be improved? Can WP be mapped? (Zwart 2010)

Increase yield, reduce ETIs it possible to reduce ET without reducing yield or to increase yield without increasing ET?

Two variables: 1. ET-biomass relationship: Light use efficiency (ε)

Not highly variable (highest for cool climates). Can efficiency be increased at a given location? 2. Harvest index (Hi) = yield/biomass

Responsible for most gains in yield in 20th centuryIf RS maps biomass, how can Hi be mapped?

APAR = absorbed photosynthetically active radiation

Van Dam et al, 2003 Sirsa irrigation district, India

Can water productivity be improved?

Difficult/impossible to improvefor a given crop, nutrients, climateShift cropping calendar to low PET months?Improve nutrient status

Reduce E from open water, soil(non-beneficial ET)

Increase harvest index = Y/DMReduce loss from pests, spoilageReduce water stress in critical periods Detectable with RS?

M2 poor nutrient statusM3 cooler climate

M4 reduce E

Is it possible to map water productivity using RS? Do we need to map ET for WP?

What is the state-of-the-art in predicting yield and production using RS?

Gap filling on cloudy days Example: Only 22% of days had cloud-free imagery in Hai Basin, China Combine microwave soil moisture (SM) with linearly interpolated LAI, r

Wu et al, 2012

Rising food prices and food security: Role of supply shocks in food security. Where did the collapse in production occur?

Interannual variability in water resources and irrigated area

Potential research focus: “What is the spatial pattern of fluctuations in irrigated area?” “Are there years where many irrigation systems globally show large decreases?” “What is the global response of irrigated area to climate fluctuation?” “Is there global synchronicity in irrigated area fluctuation?”

Link irrigated area to hydrological processes 1. ET 2. Surface water observations 3. Groundwater observations

Districts dominated by small reservoir irrigation in 1950s. Interannual variability in tanks vs gw