Conclusions - Wetlands in East Africa

1 downloads 0 Views 3MB Size Report
2, 39114 Magdeburg, Germany. Contact: [email protected] Experimental design wetland scale. First results wetland scale. Modeling concept.

Hydrological modeling on multiple scales in a data scarce catchment in East Africa Introduction

Research Area

The Kilombero floodplain in Tanzania is one of the focal areas of the “GlobE - Wetlands in East Africa” project which emphasizes on reconciling future food production with environmental protection in East African wetlands. The project covers multiple hydrological scales from the plot scale up to the regional scale in the countries Kenya, Rwanda, Tanzania and Uganda. This study concentrates on wetland-catchment interactions in Kilombero Valley which is envisaged as large scale rice production area in the forthcoming years. We are analyzing historical, current and future water availability and distribution inside the catchment and its inherent wetland.

Modeling concept

- The Kilombero wetland is characterized by a groundwater-surface water system with decreasing influence of the Kilombero river from the center to the fringe position • Different models on catchment - wetland scale: 1. Point scale: Hydrus 1D 2. Flood model: HEC-RAS 3. Hydrological models : SWAT and SWATgrid 4. Hydrochemical model: PHREEQ-C • Changing land use patterns require distributed hydrological modeling to simulate future water availability (SWAT/SWATgrid) • Integration of global change scenarios (land use and climate) on catchment scale will be conducted

- Results on plot scale show impact of soil cultivation of the upper two layers on soil physical properties and soil moisture dynamics - Climate station data is scarce in time and space on catchment scale therefore different data sources have to be used

Experimental design wetland scale

Geology & Soils

Climate & Hydrology

• part of the East African Rift System • highlands : gneiss and acid granulites • valley : sandstone and sediments covered by alluvial material • catchment: dominance of Acrisols and Ferralsols • wetland: mainly Fluvisols

• sub-humid tropical climate • ~1300 mm precipitation • unimodal and bimodal rainfall patterns • runoff coefficient ~0,38 • annual flooding • perennial flow of the Kilombero river mainly supplied by its northern tributaries

• The floodplain is affected by significant land use changes especially for rice production • A mixture of different rainfall products (station, reanalysis, satellite data) has to be applied to get the best possible spatial and temporal representation of precipitation

First results wetland scale Ground water level and daily precipitation at two different hydrological zones 2.00

0.00 20.00


40.00 1.00 60.00 0.50



100.00 120.00

-0.50 Precipitation, mm

Center position


Fringe position




Daily precipitation [mm]

- The shallow groundwater level at wetland scale responds quickly to precipitation, sub-surface flow and flooding

• lowland floodplain framed by mountains up to 2500 m a.s.l. • total catchment size: 40.240 km²

Daily Water level [m]




Workflow • Strong groundwater-surface water interaction • Mixed and spatially heteorogenous groundwater-surface water impact on soil saturation during the rainy season 0.8



1 2

0.5 0.4




Precipitation [mm/h]

Soil Water content [-]


0.2 Precipitation 30 cm_modelled 20 cm_measured

0.1 0 0


10 cm_modelled 40 cm_modelled 30 cm_measured 100

Time [h]



20 cm_modelled 10 cm_measured 40 cm_measured 200

6 250

• Application of Hydrus 1D model to simulate soil moisture dynamics at the transition period between dry and wet season at the experimental center position. • Soil moisture dynamic is controlled by soil properties and land use during the dry season as well as during the transition periods (dry-wet and wet-dry) and is controlled by flooding during the rainy season Plot scale experiments at three different hydrological regimes of the wetland transect: • agricultural field trials (paddy rice) • hydrological instrumentation (soil moisture, gw-level) • soil analysis (nutrients + physical properties)

Authors: K. Näschen 1,S. Beuel 2, B. Diekkrüger 1, G.Gabiri 1, C. Leemhuis 1, A. Müller 1, M. Ziegler3 1University

of Bonn, Department of Geography, Meckenheimer Allee 166, 53115 Bonn, Germany 2University of Bonn, Department of Geology, Nussallee 8, 53115 Bonn, Germany 3University of Applied Sciences Magdeburg-Stendal, Department of Water and Waste Management, Breitscheidstr. 2, 39114 Magdeburg, Germany Contact: [email protected]

Modelled soil water content [%]


• Modeling results were good with R² of 0.92 and NashSutcliffe Efficiency of 0.9

R² = 0.92



• The calculated error of the water balance was < 1%

0.15 0.15



Measured soil water content [%]


Suggest Documents