1. On the Causes of the Shrinking of Lake Chad
Huilin Gao1, Theodore Bohn1, Erika Podest2, Dennis P. Lettenmaier1
1Dept. of Civil and Environmental Engineering, University of Washington
2Jet Propulsion Laboratory, California Institute of Technology

2. Approach Description of the physical lake system
Hydrological model and its implementation
Historical reconstruction of Lake Chad using the model
Evaluation of irrigation effect versus climate effect
Role of lake bathymetry
Evaluate prospects for natural recovery and effects of water diversion

3. Lake Chad basin and Lake Chad
Largest endoreic basin in the world
Fourth largest lake in Africa 40 years ago
2,500,000 km2

4. The unique bathymetry of Lake Chad
River Chari provides 98% of the river inflow into Lake Chad from the south;
284
283
282
281
280
279
278
277
276
When water is effluent/deep, the lake behaves as one lake;
When water level retreats below the barrier, the lake splits into two parts.
Bol (gauge)
Great barrier
Inflow
(m)
Inflow

5. The loss of Lake Chad: what happened?
Can we quantify these by hydrological modeling? ?
Human water usage? Lake bathymetry?
Climate change?
10/31/1963
12/25/1972
01/31/1987
02/06/2007
precipitation

6. Variable Infiltration Capacity (VIC) model
Semi-distributed model driven by a set of surface meteorological data
Represents vegetation, has three soil layers with variable infiltration, non linear base flow
Simulated hydrology at each grid cell for all time steps
Well calibrated and applied in a number of large river basins over the continental US and the globe.

7. VIC lake algorithm
I: Evaporation from the lake is calculated via energy balance;
II: Runoff enters the lake from the land surface;
III: Runoff out of the lake is calculated based on the new stage;
IV: The stage is re-calculated.
(Bowling and Lettenmaier, JHM, 2010)

8. Modeling approach Simulate over the drainage basin
Calibrate model parameters using observed discharge from 1952 to1963;
Run the model from 1952 to 2006.
2. Implement the model to represent Lake Chad dynamics
Switch between one-lake and two-lakes according to lake level;
Add the inflow as additional forcing.
Inflow

9. Validation of modeled lake depth
(observations from gauge and satellite altimetry) 12 10 8 6 4 2
Gauge
(RMSE=0.37m)
One lake
Lake depth (m)

10. Validation of modeled lake depth
(observations from gauge and satellite altimetry) 12 10 8 6 4 2
Gauge
(RMSE=0.37m)
Satellite
(RMSE=0.48m)
One lake
Lake depth (m)
south lake
north lake

11. Validation of modeled lake surface area
(observations from satellite imagery)
Landsat/aircraft derived
water coverage
VIC simulated
lake area
Landsat/aircraft images
Bias: -8%
10/31/1963
Bias: 3%
12/25/1972
Bias: 5%
01/31/1987

12. Irrigation impact Lake depth (m) 12 10 8 6 4 2 not Irrigated
Lake depth (m)
not Irrigated
(south lake)
not Irrigated
(north lake)

13. Climate impact
An equilibrium of two lakes is favored by climatology

14. Climate impact
The threshold for an equilibrium of one lake is 38km3/yr inflow

15. Bathymetry impact Lake depth (m) 12 10 8 6 4 2
Simulation with the barrier removed
no historical split
Lake depth (m)
irrigated
If the lake did not split, Lake Chad still would have shrunk dramatically

16. Irrigation vs split: lake volume
Irrigation withdrawals played a key role in preventing a recovery
Lake bathymetry exacerbated the loss of the lake

17. Irrigation vs split: lake volume
Lake volume would be maximized if there was neither irrigation nor split

18. Towards recovering Maximum inflow from 1952 to 2006: 50 km3/yr
Observed inflow from 1997 to 2006:
24 km3/yr
Naturalized inflow from 1997 to 2006:
37 km3/yr
Inter-basin water transfer is needed for a full recover of Lake Chad
1963 level
Merge level

19. Conclusions
The VIC model simulated lake dynamics are very consistent with gauge data and satellite observations;
The bifurcation of Lake Chad in 1972 occurred as a combined consequence of the bathymetry of the lake and severe droughts;
Absent irrigation, the lake would have merged around 2000;
The averaged climatology from 1952 to 2006 does not favor a single lake;
Supplemental water transfer is required for the lake to recover to its 1963 size.

20. Thanks! On the Causes of the Shrinking of Lake Chad
Huilin Gao1, Theodore Bohn1, Erika Podest2, Dennis P. Lettenmaier1
1Dept. of Civil and Environmental Engineering, University of Washington
2Jet Propulsion Laboratory, California Institute of Technology
Thanks!

21. Modeling strategy and results
Analyzing Lake Chad from a hydrological perspective ……
10/31/1963
12/25/1972
01/31/1987
02/06/2007
i) Validate the modeled lake dynamics ( ) using gauge and satellite observations;
- How well could the model capture the loss of Lake Chad?
ii) A set of experiments to test the impacts due to climate change, human water usage, and lake bathymetry;
- How much do these factors affect Lake Chad?
iii) What it takes for a full recovery for the lake.
- Will Lake Chad be able to recover naturally?

22. VIC wetland algorithm
a) when the lake is at its maximum extent the soil column is saturated;
b) as the lake shrinks runoff from the land surface enters the lake;
c) evaporation from the land surface depletes soil moisture;
d) as the lake grows, water from the lake recharges the wetland soil moisture.
(Bowling and Lettenmaier, JHM, 2010)

24. iii) Irrigation vs split: lake area & volume

25. iii) Irrigation impact: lake depth12 10 8 6 4 2
Small impacts:
Delayed dry-out in north lake;
Increased lake level in south lake.
not irrigated
(south lake)
historical split
Lake depth (m)
Irrigated
(south lake)
not irrigated
(north lake)
Irrigated
(north lake)

26. i) Validation of modeled lake depth
(observations from gauge and satellite altimetry) 12 10 8 6 4 2
gauge
One lake
Lake depth (m)
south lake
north lake 26

27. ii) Split/bathymetry impact12 10 8 6 4 2
Irrigation impact would be magnified if the historical split of the lake did not occur
not irrigated
no historical split
Lake depth (m)
irrigated