Observation of Sensible and Latent Heat Fluxes Over a Mixed Savanna – Agricultural Catchment in the Semi-Arid Sudanian Savanna, Burkina Faso

Evaporation and the surface energy balance over two land cover types in Sudanian savanna were investigated in terms of their seasonal and interannual variability using four years of eddy covariance measurement (2009-2013). The observations demonstrate the strong controls on evaporation by vegetation through modification of the surface roughness, albedo, and available moisture related to seasonal variations in land cover and topography. Two characteristic land uses of semi-arid West Africa (Burkina Faso) were monitored: agricultural fields and gallery forest. The sites receive around 800 mm of rain most years, typically between the months of May and October, with up to approximately 20% being transferred to runoff. Seasonal variation in all components of the energy balance was found to be greater over the agricultural landscape, the total latent energy flux over the gallery forest was higher by approximately 30%, and the evaporative fraction was more dependent on soil moisture in the agricultural fields. According to the diurnal behavior of the evaporative fraction, we isolate constant periods that we compare with measures of cloudy sky, NDVI and soil moisture. We identify predictors of diurnal variation of the evaporative fraction according to land cover, which allows incorporation of remote sensing and satellite data. Improved understanding of the variability of fluxes over diverse land cover will allow us to improve the estimation of evaporation to the atmosphere over the entire watershed. Ongoing outreach projects accompany this research to integrate findings with local knowledge to improve resilience of small-scale rain fed agriculture.

Evaporation from cultivated and semi-wild Sudanian Savanna in west Africa

Natalie Claire Ceperley, Theophile Mande, Marc Parlange, Nicolaas Cornelis van de Giesen

Rain-fed farming is the primary livelihood of semi-arid west Africa. Changes in land cover have the potential to affect precipitation, the critical resource for production. Turbulent flux measurements from two eddy-covariance towers and additional observations from a dense network of small, wireless meteorological stations combine to relate land cover (savanna forest and agriculture) to evaporation in a small (3.5 km(2)) catchment in Burkina Faso, west Africa. We observe larger sensible and latent heat fluxes over the savanna forest in the headwater area relative to the agricultural section of the watershed all year. Higher fluxes above the savanna forest are attributed to the greater number of exposed rocks and trees and the higher productivity of the forest compared to rain-fed, hand-farmed agricultural fields. Vegetation cover and soil moisture are found to be primary controls of the evaporative fraction. Satellite-derived vegetation index (NDVI) and soil moisture are determined to be good predictors of evaporative fraction, as indicators of the physical basis of evaporation. Our measurements provide an estimator that can be used to derive evaporative fraction when only NDVI is available. Such large-scale estimates of evaporative fraction from remotely sensed data are valuable where groundbased measurements are lacking, which is the case across the African continent and many other semi-arid areas. Evaporative fraction estimates can be combined, for example, with sensible heat from measurements of temperature variance, to provide an estimate of evaporation when only minimal meteorological measurements are available in remote regions of the world. These findings reinforce local cultural beliefs of the importance of forest fragments for climate regulation and may provide support to local decision makers and rural farmers in the maintenance of the forest areas.

European Geosciences Union2017

Hydrology of the Sudanian Savannah in West Africa, Burkina Faso

Theophile Mande

As in all semi-arid regions, Tambarga, a small village surrounded by national parks in the landlocked country of Burkina Faso, is affected by the seasonality of the local hydrology. Seasonal and spatial variability of rainfall shapes the livelihoods of rural farmers who depend mainly on rain-fed agriculture. We instrumented the Tambarga catchment (area=~4 km2) to investigate the rainfall patterns, the resulting hydrologic processes (surface runoff, base flow) and the evaporation effects on the local water balance. Thus, we have measured hydrological, meteorological and soil parameters at high spatial and temporal resolution over the catchment since 2009. Data for this research were acquired from the 2010 to 2012 rainy seasons using a network of automatic wireless weather stations (SensorScope stations), two weirs, eight piezometric wells, and two surface energy balance stations. Rain at Tambarga is triggered by the convective mechanism, which is mainly controlled by the sensible heat flux. The daytime rain events at Tambarga are convective and are characterized by their high intensities and short durations. An overall increase in rainfall of 10-30% is observed in the savannah forest when compared to the agricultural field from 2010 to 2012. Biotic (leaf area index) and abiotic (well drained soil) factors, causing enhanced sensible heat flux at the savannah site, lead to an increased predisposition toward convective rainfall. The hydrologic processes concomitant to the rain events were identified over the basin and intermittent responses of streamflow were isolated. The rise of the perched water tables located in the upstream and downstream regions of the basin generated two separate flows in the riverbed after the first rain. The intermediate zone operates like a deep storage tank, and by filling, it creates a connection between the two perched water tables. This connection induces a continuous flow throughout the 2.8 km of the stream. The hydrologic response at the basin outlet is observed to alternate between two different states over the course of the season. At the start of the rainy season, when the soils are very dry and the groundwater level is deep, a typical single-peak hydrograph is observed. This evolves into a double-peak hydrograph when the rainy season is completely established. The single-peak hydrograph is correlated with rainfall intensity, while the double-peak hydrograph is also correlated with the antecedent soil moisture condition. The baseflow, meanwhile, occurs when the groundwater level is higher than the riverbed at certain locations in the basin. In contrast, during dry days (no rain) the baseflow exhibits a diurnal flow pattern. This diurnal pattern consists of decreased flow rate from sunrise (6:00 A.M.) to 1:00 P.M. and flow rate recovery to its previous level from 1:00 P.M. to sunset (6:00 P.M.). The diurnal pattern of the streamflow was found to be derived both from infiltration in the riverbed and evaporation over the river edges. The infiltration process dominates during the beginning of the season. While late in the season, when the groundwater network is interconnected and its overall level is higher than the riverbed elevation, evaporation controls the diurnal pattern. An evaporation contributing area is defined when the diurnal pattern is controlled by the evaporation. This area is about 0.6% of the basin area and could be rationally represented by the riparian area and the outlet wetland. Given the importance of rainfall-runoff processes over this catchment, a water balance was completed based on a simple lumped model. This model allowed for testing some water management strategies to improve agricultural production. Its implementation suggests that storing water underground for irrigation purpose with deeper wells is a way to achieve better agricultural productivity and therefore, reduce the vulnerability of the local people.

EPFL2013

Water use by Sclerocarya birrea Agroforestry Trees in Sudanian Savanna: Application of Wireless Sensor Networks

Natalie Claire Ceperley

Small scale rain fed agriculture is the primary livelihood for a large part of the population of Burkina Faso. Regional climate change means that this population is becoming increasingly vulnerable. Additionally, as natural savanna is converted for agriculture, hydrological systems are observed to become less stable as inﬁltration is decreased and rapid runoff is increased to the detriment of crop productivity, downstream populations and local water sources. The majority of the Singou River Basin, located in South East Burkina Faso is managed by hunting reserves, geared to maintaining high populations of wild game; however, residents surrounding the protected areas have been forced to intensify agriculture that has resulted in soil degradation as well as increases in the frequency and severity of ﬂooding and droughts. Agroforestry, or planting trees in cultivated ﬁelds, has been proposed as a solution to help buffer these negative consequences, however the speciﬁc hydrologic behavior of the watershed land cover is unknown. Sclerocarya birrea, or Marula, is a dominant agroforestry tree in the region valued for its medicinal, ritual, and hydrologic services. We explore the effect of Sclerocarya birrea trees on the soil moisture response to precipitation in millet-dominated agricultural ﬁelds in the village of Tambarga, Burkina Faso. We compare the water budget constructed surrounding a single Sclerocarya birrea tree with that in an open ﬁeld using measured values of soil water content, canopy inﬁltration, precipitation, evaporation. Data are collected through a distributed sensor network of 17 Sensorscope wireless meteorological stations in addition to 2 eddy covariance stations and sap ﬂow probes. These stations are dispersed across cultivated rice and millet ﬁelds, natural savanna, fallow ﬁelds, and around agroforestry ﬁelds. Three of the stations are uniquely conﬁgured around Sclerocarya birrea trees. Sensorscope routes data through the network of stations to be delivered by a GPRS connection to a main server. Data are available in real time via a website that can be accessed by a mobile phone. The stations are powered autonomously by small photovoltaic panels. This deployment is the ﬁrst time that these meteorological stations have been used on the African continent. Results from this research are being coupled with outreach activities in the community in order to best integrate them into local practices. Our hope is that the results of our modeling will inform local farmers and thus help improve their adaption to changing weather patterns and land cover.

2011