Impacts of shade trees on hydrological services and erosionin a coffee AFS of Costa Rica: scaling from plot to watershed

Federico Gómez-DelgadoPhD Student

Roger MoussaBruno RapidelRintaro KinoshitaAlexis PerezPhilippe VaastJean-Michel HarmandOlivier Roupsard

CAFNETProjectINRA

Research questionsWhat is the rate and type of erosion in coffee watersheds and how much do trees reduce it ?

we propose to assess and model the water balance and sediment yield from plot to watershed

Can AF practices be promoted and rewarded as hydrological services?

Hydropower profitability is limited by sedimentation of reservoirs. The tools here developed could help in the negotiation of hydrological services between coffee producers and hydropower institutes/companies in Costa Rica

Study site: Cafnet/CoffeeFlux experiment

Aquiares experimental watershedArea = 1 km2

Mean slope: 20%Max. slope: 89%

Turrialba river watershed

Area = 77 km2

Coffee plantations

Turrialba volcano

COSTA RICACENTRAL AMERICA

Cafnet / CoffeeFluxExperimental display CO2

H2O

Vapor, Carbon, Climate

Flux Tower

• LAI• Interception• Throughfall• Stemflow• Sapflow

Plants + Treesflow experiments

Soil Tubes

Soil water content

Rainfall Stations

Rainfall Streamflow + Turbidity

Hydraulic Flume

Water table level

Piezometers

Experimental Plots

S.Runoff + Erosion

• Infiltrability• Hydraulic conductivity

Soil properties experiments

Rainfall and streamflow

Energy balance, ET flux

Cumulative water balance

Potential and actual ET (AET)

Soil moisture (SM)

Water storage: upper GW box (STZ) middle GW box (SUZ)

Water storage: lower GW box (SLZ)

Observed precipitation (P)

Observed and simul. streamflow (Q)

Simulation of water balance HBV lumped model

very high infiltration => low runoff~> low laminar erosionthen: flooding events non-laminar erosion

Sediment loss by watershed

Surface Runoff (SR) and Sediment Yield (SY)in shaded and non-shaded experimental plots

Experimental Plots

Water balanceEvent 29/06/09, watershed scale

Sediment Yield (SY)Event 29/06/09, plot to watershed scale

Total SY = 564 kg (100%)(measured at theoutlet catchment)

- SY non-shade = 169 kg (30%) - SY shade = 51 kg (9% ) SY road+n.lam.= 344 kg (61%)

Scaled upScaled up

By diff.

2/3 of watershed sediment yield comes from non-laminar sources!

Conclusions

We designed an experimental display in a coffe AF system

(andosols, montaneaous) in order to assess and model:

· the water balance partitioning from plot to watershed

· the sediment yield from plot to watershed

A combination of techiques and models allowed to evidence that:· streamflow was dominated by the base flow (aquifer)

· the infiltration rate was enormous in these conditions (andosols, many roots)

· the laminar erosion was low as compared to the non-laminar erosion

Perspectives

• to close water balance on one full year at least

• to continue accumulating sediment yield

• to assess the influence of tree density and slope on plot erosion and scale up to watershed

• to explain the origin of non-laminar erosion and see how management can reduce it

• to validate in other contrasting watershed (other soil, other slope, other tree density...)

Many thanks!

End of presentation