Reconstruction of Inundation and Greenhouse Gas Emissions from Siberian Wetlands over the

Last 60 Years

T.J. Bohn1, E. Podest2, R. Schroeder2, K.C. McDonald2, L. C. Bowling3, and D.P. Lettenmaier1

1Dept. of Civil and Environmental Engineering, University of Washington, Seattle, WA, USA

2JPL-NASA, Pasadena, CA, USA,3Purdue University, West Lafayette, IN, USA

Steve Burges Retirement SymposiumUniversity of WashingtonSeattle, WA, 2010-Mar-24

West SiberianLowlands

(Gorham, 1991)

Western Siberian Wetlands

30% of world’s wetlands are in N. Eurasia

High latitudes experiencing pronounced climate change

Response to future climate change uncertain

Wetlands:Largest natural global source of CH4

Climate Factors

Water Table

Living Biomass

Acrotelm

Catotelm

Aerobic Rh

CO2

Anaerobic Rh

CH4

Precipitation

Temperature(via evaporation)

Temperature(via metabolic rates)

NPP

CO2

Water table depth not uniformacross landscape - heterogeneous

Relationships non-linear

Note: currently not considering export of DOC from soils

Modeling Framework

• VIC hydrology model– Large, “flat” grid cells (e.g.

100x100 km)– On hourly time step,

simulate:• Soil T profile• Water table depth ZWT

• NPP• Soil Respiration• Other hydrologic variables…

• Link to CH4/CO2 emissions model (Walter & Heimann 2000)

How to represent spatial heterogeneity of water table depth?

Distributed Water Table

0 1CumulativeArea Fraction

кi

кmax

кmin

Topographic Wetness Index CDF

Topographic WetnessIndex к(x,y)

DEM (e.g. GTOPO30or SRTM)

0 1CumulativeArea Fraction

Water Table

SaturatedSoil

Soil Storage Capacity CDF(mm) = f(кi)

Smax

0

VIC SoilMoisture (t)

Water Table Depth Zwt(t,x,y)

Resolution = 1 km

Summarize for One100km VIC grid cell

High-Resolution Lake ObservationsNeed multi-temporal observations

Remote Sensing products:(courtesy JPL Carbon and Water Cycles Group

(E. Podest, R. Schroeder, and K. McDonald))

• LANDSAT open water classifications– 30m– Repeat cycle: decadal– 1987-2007

• PALSAR open water, inundation, and saturated soil– 30m– Repeat cycle: sporadic– 2006-2007

• AMSR inundation– 25km– Repeat cycle: 10 days– 2006-2007

1 km Landcover Classification(Bartalev et al 2003): 0.1% Open Water30m LANDSAT Open Water Class.1989-08-28: 1.1% Open Water30m LANDSAT Open Water Class.1999-09-16: 2.4% Open Water30m LANDSAT Open Water Class.2006-07-01: 4.1% Open Water

Study Domain: W. Siberia

Wetness Index from GTOPO-30 and SRTM3

Ural Mtns Yenisei R.

Ob’ R.

Vasuygan Wetlands

Chaya/Bakchar/Iksa Basin

Close correspondence between:

•wetness index distribution and

•observed inundation of wetlands from satellite observations

GLWD Wetland delineation(Lehner and Doll, 2004)

Chaya/Bakchar/Iksa Basin

100km Grid Cells

Close correspondence between wetness index and wetland vegetation

4 focus areas (see next slide)

Comparison with PALSAR

Observed Inundated Fraction (PALSAR Classification)

Simulated Inundated Fraction (at optimal Zwt)

ROI 12006-06-09

ROI 22007-07-06

ROI 32006-05-28

ROI 42007-07-18

Observed Inundated Fraction (PALSAR Classification)

Simulated Inundated Fraction (at optimal Zwt)

•Spatial distribution of inundation compares favorably with remote sensing•This offers a method to calibrate model soil parameters

Approx.30 km

How do resulting emissions differ between uniform water table and distributed water table?

Experiment:• Calibrate methane model to match

in situ emissions at a point (Bakchar site, Friborg et al, 2003)

• Distributed case: calibrate distributed model water table depth to match observed inundation

Water Table

CH4

• Uniform case: select water table timeseries from single point in the landscape having same long-term average methane emissions as the entire grid cell in the distributed case; apply this water table to entire grid cell

Inundated Area(matching remote sensing)

Interannual Variability, 1948-2007

Uniform Water TableDistrib Water Table

Uniform Water Table:

Shallower than average of distributed case

But never reaches surface; no inundation

Resulting CH4 has higher variability than for distributed case

Distributed case is buffered by high- and low-emitting regions

Possible trend in temperature, also in CH4

Impact on trends?

Net Greenhouse Warming Potential

NPP

RhCO2

RhCH4

RhCO2 - NPP

NET GHWP

NPP and RhCO2 approximately cancel

Net GHWP essentially follows GHWP(CH4)

Uniform water table:

•CH4 has larger interannual variability

•So does net GHWP

•Impact on trend assessment?

CH4 makes up small part of C budget, but large contribution to greenhouse warming potential

On 100-year timescale, GHWP(CH4) = approx. 23 * GHWP(CO2)

Conclusions

• Advantages of distributed water table:– Facilitates comparison with satellite

measurements and point measurements– More realistic representation of hydrologic

and carbon processes

• Spatial distribution of water table has large effect on estimates of greenhouse gas emissions and their trends

Thank You

This work was carried out at the University of Washington and the Jet Propulsion Laboratory under contract from the National Aeronautics and Space Administration.

This work was funded by NASA grant NNX08AH97G.

Calibration – Bakchar Bog, 1999

Soil T

ZWT (water table depth)

CH4

VBM = VIC-BETHY-Methane(Bohn et al., 2007)