Evaluation,of,Surface,Water,Availability,for,Inland,Valleys,Rice,Production:,The,Case,of,Mankran,Wat for和of的用法

　　Busia N. Dawuni1, Regassa Namara2, Fred Kizito2 and Hideto Fujii3 　　1. Ghana Irrigation Development Authority, Accra, Ghana
　　2. International Water Management Institute, Accra, Ghana
　　3. Japan International Cooperation Centre for Agricultural Sciences (JIRCAS), Ibaraki 305-8686, Japan
　　Received: May 20, 2011 / Accepted: July 22, 2011 / Published: February 20, 2012.
　　Abstract: In Ghana, inland valleys have been found to be suitable for rice cultivation and could potentially increase lowland paddy rice production. This study estimated the temporal variations of surface water resources and their spatial distribution in the Mankran watershed of Ghana through the collection of recorded hydrological data in the bench mark watershedfrom2008 to 2010. Since most inland valley rice cultivation highly depends on rainfall, the watershed precipitation data over a decadal period showed sufficient wet years with a potential to sustain a high cultivation of paddy rice. Peak wetness occurred in June and October over a bi-modal precipitation pattern. Rating curve data for the Mankran-kesse river-upstream depicted low discharge values despite having a higher stream order. Thus stream order alone was not sufficient to estimate water resources potential. It was presumed that the geomorphology and lithology of the highly porous river bed and the presence of high sub-surface water resources stored in this zone may be implicated for this observation. Provision of water storage options for zones like Kesse-upstream seems a feasible option in order to cater for supplementary irrigation while indirectly tapping on subsurface water resources stored in the porous aquifers through basin interflows. Base flow data also showed that the discharge from upstream locations to the downstream exit of the watershed was high through direct surface river discharge and subsurface interflow. The temporal patterns of the hydrology indicate that annual paddy rice cultivation is ideal between May and October.
　　Key words: Mankran watershed, inland valleys, surface water availability, rice production.
　　 1. Introduction??
　　The world’s population is growing fast and is projected to hit the 9.1 billion mark by 2050 [1]. The challenge therefore is how to produce enough to meet the growing demand for food and fiber. It is estimated that 925 million people all over the world are suffering from malnutrition, of which a large proportion are found in developing countries. In order to address this deficit and ensure adequate food for the projected increase in population by 2050, food production has to increase by 70% globally and double in the case of developing countries [1].
　　In sub-Saharan Africa, about 30% of the population is undernourished. Food and rural development research indicates that unless food production is doubled in the next 25 years, the famine situation will worsen [2]. Rice is increasingly becoming a major staple in West Africa including Ghana which may be the result of urbanization and the changing taste of the rural folk. Previously, rice was consumed mostly by the urban dwellers and by the rural folk only during festive occasions but currently, this scenario does not hold true.
　　Statistics from the African Rice Center show that rice consumption in Sub-Saharan Africa (SSA) including Ghana has consistently grown since the 1960s to date. In Ghana, the per capital consumption of rice per annum has increased from 12.7 kg in 1985 to 15.1 kg in 2005 [3]. However, the increase in production has not been able to match the increase in consumption. As a result, Ghana is highly reliant on imported rice to compensate for the local deficit. This has substantially contributed to depleting the meager foreign exchange base. It is estimated that the consumptive requirement of rice in 2008 was about 500,000 metric tons of which local production was only able to meet 150,000 metric tons. In the year 2004, about 253,905 metric tons of rice were imported. This quantity increased to about 389,660 metric tons in 2006. In terms of foreign exchange, 119.15 million and 159.47 million US dollars were spent in 2004 and 2006 respectively [3]. It has been estimated that if 2 million ha of inland valleys were under lowland rice production and yielding 3 t/ha, rice importation in West Africa could be halted [4]. The need to increase paddy rice production in Ghana is to meet the local consumptive need and also to reduce the foreign exchange burdenis therefore paramount. Two major issues suggested by FAO to increase food production in West Africa are: (1) the expansion of rainfed cultivated areas; (2) the increase in both large and small scale irrigation [1]. There is however a limited scope for the expansion of rainfed areas whilst increase in irrigation area is hampered by high infrastructure developmental cost which highlights the importance of rice cultivation in inland valleys [1].
　　Inland Valleys Rice Production and Study Objectives: Rice is a hydrophilic and semi-aquatic plant and is usually grown under wetland conditions. Under favorable hydrological conditions and with proper on-farm water management and cultural practices, there is potential to have high rice yields under inland valleys paddy production. In Ghana, inland valleys have been found to be suitable for rice cultivation and therefore could potentially increase lowland paddy rice production. Under the “sawah”technology (Bunded field for paddy cultivation in inland valleys) for semi-irrigated rice in Ghana, paddy yields have increased four times in inland valleys where these technologies are being practiced around Mankransoin the Ahafo-Ano South district of the Ashanti region [5].
　　Ironically, inland valleys in Ghana have not been fully exploited. Research in inland valleys cultivation in Ghana has concentrated more on agronomic practices and soil characteristics with knowledge gaps on their hydrological characteristics. Stream flows in these inland valleys are a good water resource for lowland paddy production. However, very little data exist on their flow regime and discharges. In addition, the surface water resource potential of most of the valleys is not known. Knowledge of the volumes of flow and spatial distribution is essential for the evaluation of the suitability of the valleys for paddy cultivation from the surface water resource potential viewpoint. Furthermore, sustainable development areas could also be estimated.
　　The Mankran watershed is composed of numerous inland valleys with portions currently being developed by the Inland Valleys Rice Development Project(IVRDP) of the Ministry of Food and Agriculture(MoFA) for paddy rice cultivation. However, data on stream flow is scanty and insufficient to serve as a guide to any effective lowland rice developmental program which necessitated this study. The main objective of this study was to estimate the temporal variations in surface water resources and their spatial distribution in the Mankran watershed for lowland paddy rice production under semi-irrigated conditions.
　　 2. Study Area
　　The study area is the Mankran watershed which is located in the Ahafo-Anosouth district of the Ashanti region, Ghana. It is about 45 km to the north-west of Kumasi (Fig. 1), and lies in an equatorial rain forest zone with a catchment area of about 500 km2. The average annual rainfall at the benchmark site is 1,450 mm and is bimodal with the first peak occurring in June whereas the second peak occurs in October. The Mankran watershed is composed of foursub-watersheds namely: Mankrankesse upstream(Kesse), Biem, Dwinyanand Makrankuma (Kuma). The drainage system in the basin is composed of a network of rivers whose flows vary from perennial to ephemeral with a drainage density of 1.2-2.4 km/km2[6]. The Mankrankesse is the main river with a good number of tributaries. Other big rivers are the Biem and Mankrankuma.
　　 3. Methodology
　　The study was conducted through the collection of recorded hydro-meteorological data (rainfall and stream flow) in the bench mark watershed. This included point rainfall from the Mankranso meteorological station for a period of 10 years from 1997 to 2006 (Fig. 2). In order to estimate the volume of runoff that contributed to the flow of the rivers and streams in the watershed during the study period, five rain gauges were distributed and installed over the entire catchment in five locations to record rainfall. The villages with the rainfall recorders and their designated location names are as follows: Biemso No. 1 (Ra), Kunsu (Rb), Barniekrom (Rc), Fawoman (RD), and Amanin (Re) (Fig. 1). Data were collected for a period of two years, 9 months spanning (February 2008 to November 2010). Three stations along the Mankran-kesse were selected for water level observations and these include one upstream: Mankran-kesse river-upstream-Kase(station E), one station midstream: Mankran-kesse river-midstream (station A) and another downstream: Mankran-kesse river-mainstream (Station C). The streams were classified according to the Strahler classification [7] as stream orders 4, 4 and 5 respectively. In this system, the smallest head-water
　　? tributaries are called first-order streams. Where two first-order streams meet, a second-order stream is created and where two second-order streams meet, a third-order stream is created, and so on. In addition, the order of a stream does not change when it meets with one of lower order i.e. a third order meeting a second order still remains a third order. Along the Mankrankuma River which is of stream order 4, one station designated B was also selected. At other sub-watersheds, one point along each of the three rivers was selected i.e. the Biem, Dwinyan and Nyasiso. Their respective gauging stations are at D, F and G (Fig. 1).
　　With regards to stream flow data, daily water levels were monitored at these seven flow gauging stations in the watershed. The rivers were gauged using staff gauges and designated as follows: Mankrankesse river-midstream (Station A), Mankrankuma river(Station B), Mankrankesse river-Mainstream(StationC), Biemriver (Station D), Mankrankesse river-upstream (Station E), Dwinyan River (Station F) and Nyasi River (Station G). These stations as designated are shown in Fig. 1. Water level data in the gauging stations A, B, C, D, E, F and G were collected for a period of 3 years i.e. November 2007-October 2010.
　　In order to relate the water level measurements and the discharge of the rivers, electro-magnetic current meters, the SEBA MDS-Surfloat dataloggers(SEBA-Hydrometrie, Germany, Inc.) were used as integrated float-driven encoders for measurement and storage of water level fluctuations for the various streams in the watershed. These loggers recorded flow measurements and determined the velocities of flow for the various rivers at all the gauging stations except G. The field measurements were carried out six times at each gauging station. Additionally, data was collected on the cross-sectional areas where the measurements were carried out. With the measured velocity and cross sectional area determined, the flow discharge was computed.
　　The computed discharge data and the corresponding water levels measured during the survey permitted the establishment of stage-discharge (H-Q) relationship curves for all stations where measurements had been conducted (Fig. 3; Table 1). Using the derived discharge equations, the water level readings were then converted to stream discharges of rivers at the various stations (Fig. 4). For the analysis of flow through individual rivers, time series hydrographs of the rivers over the three-year periods were plotted (Fig. 4). One of the parameters used for the evaluation of the water resource potential at the sub-watershed level was the specific discharge which is the discharge per unit area of the catchment. This was computed for each sub-basin and comparison was conducted using graphical time series (Fig. 5). The estimation of surface runoff and baseflow contribution to the water resources potential was conducted using recursive digital filters (Fig. 6). (Further details are provided in the “Results and Discussion” Section).
　　 4. Results and Discussion
　　 4.1 Rainfall
　　Data obtained from the Ghana Meteorological Services Agency for the Mankran watershed is shown in Fig. 2. Two main rainy seasons are evident with the first starting in April and peaking in June and the second starting in August and having its peak in September. The minimum annual rain amount usually occurs between December and January. Annual trends for the duration 1997 through 2006 showed that the year 2006 had the highest amount of rainfall of about 1600mm whilst the lowest occurred in 1997 and was about 800mm. Inland valley rice cultivation highly depends on the rainfall amount with the hydrological wet years having a greater potential to sustain a high cultivation of paddy rice than exceptionally dry years. The rainfall standard deviation was highest (199.8 mm) in the wettest month (June) and lowest (22.4 mm) in the driest month (January). This suggests a high degree of variation of wet events over the ten-year period yet drier months depicted less deviation (Fig. 2).
　　4.2 Stage-Discharge Curves
　　Stage-discharge curves relate the water level of a given stream and its corresponding discharge. As noted previously, in order to relate the observed water level recordings to river discharge, stage-discharge relation curves were established using data obtained from field discharge measurements. Fig. 3 shows a plot of the stage-discharge curves for the rivers in the watershed at the various gauging stations as follows: Kesse (A), Kesse-Up (E), Kuma (B), Biem (D) and Dwinyan (F) rivers.
　　The corresponding flow rating equations are as shown in Table 1and are of a quadratic function nature. For the establishment of the rating curves, five data sets were collected. For the case of gauging stations A and D, measurements could not be carried out at lower water levels as can be seen in Fig. 3 and as such the rating curves are most suitable for higher water levels above 0.5 m. However, extrapolation was carried out to determine the discharge at these low water levels. It can also be observed that the rating curve for station E shows generally low values of discharge though the stream at gauging station E has a stream order of 4. This could be attributed to the geomorphology and lithology of the river bed which is highly porous. It further suggests presence of high levels of sub-surface water resources storage for this area. 4.3 Hydrograhs
　　For gauging stations A and B, the maximum discharge of the rivers during the three year period was 7.36 m3·s-1 and 11.92 m3·s-1 respectively whilst their average discharges are 0.81 m3·s-1 and 0.65 m3·s-1 respectively. The respective average discharges of other gauging stations are as follows D: 0.32m3·s-1; E: 0.12 m3·s-1; F: 0.14 m3·s-1 and the maximum discharges are: D: 2.3 m3·s-1; E: 1.14 m3·s-1; and F: 3.51 m3·s-1. As shown in Fig. 4, it can be observed that the river flows are substantial between May through October and corresponds well with the precipitation patterns. The hydrographs depict that the Kuma (B) river has a more stable and relative uniform flow during the river flow period. In the case of the Biemriver (Station D), though the flow is substantial, there are high fluctuations and as such a weir facility is necessary for the diversion of water.
　　Based on field experience, an average flow of about 0.324 m3·s-1 is adequate for the cultivation of paddy if this flow can be sustained for at least three months. The river flow characteristics of the Kesse-upstream rather shows a minimal amount of flow with an average of only about 0.12 m3·s-1 though the river has a stream order of 4 with a similar stream order as the Biem river (D). This may as well be attributed to the river basin geomorphology and the lithology of the river bed which probably allows flush flow and slow release of subsurface water stored in deep porous aquifers. The flow of the Dwinyan river at F is also quite minimal and averaged at only 0.14 m3·s-1. In the case of the Kesse-upstream and Dwinyan rivers, F, providing water storage options such as small reservoirs may be a feasible option for supplementary irrigation. The discharge at the exit of the watershed was quite substantial and shows that vast water resources are stored within the subsurface subsequently draining to the downstream exit of the watershed either through direct surface water river discharge or subsurface interflow.
　　4.4 Comparison of Discharge among Sub-Watershed in the Mankran Watershed
　　The sub-watersheds in the Mankran watershed have distinctive river flow characteristics. Discharge measurements were carried out in the sub-basins and corresponding unit hydrographs (i.e. discharge/catchment area) were determined in order to evaluate the surface water potential at sub-basin level.
　　Fig. 4 is a graph that shows the hydrographs of the various sub-basins giving a comparison between the Kesse, Kuma, Dwinyan, Biem, Kesse-Up and Mankran sub-basins. As shown in Fig. 4, the sub-basins respond as expected to precipitation events but with varying degrees. Both the downstream exit of the watershed at Mankran and Kuma sub-basins had the highest discharge while Kesse upstream sub-basin depicted the lowest discharge (Fig. 4).
　　The specific discharge for the surface water hydrology of the Mankran watershed can be defined as the discharge per unit area of an upstream sub-watershed. In terms of individual streams, it is the discharge per unit width of stream channel. By nature of their soil attributes and topographic position in the landscape, inland valleys slowly release subsurface water which gradually contributes to the overall stream flow. The average specific discharges of the sub-basins from November 2007 to August 2009 in liters per second per ha (l·s-1·ha-1) were as follows: Mankran-kuma (B): 5.82; Biem (D): 4.65; Mankrankesse-Up (E): 1.14; Dwinyan (F): 3.75. Station B has the highest average specific discharge whist E has the lowest of only 1.14 l·s-1·ha-1 even though the river is of the 4th order. The extremely low value may be attributed to the basin characteristics especially the river bed lithology. It is possible that a lot of the water in the sub-basin goes into ground water reducing the volume of surface water before reaching the flow gauging station. Another reason could be attributed to insufficient runoff generated from the upper catchment of the basin. Paddy cultivation in the area of Mankrakesse-Up (E) may not be suitable without supplementary water storage systems. There is a general observation for each year which begins with low flow, and spikes between June and October with most peaks occurring around October then returning to a low flow for the remainder of the year.
　　Table 2 gives an analysis of the surface water potential in relation to the specific discharge and stream order for the sub-basins from November, 2007 through October, 2010. Based on Table 2 and Fig. 5, it can be observed that within the limits of the data recorded, the Mankran-kuma shows the highest surface water potential with a rating of 1 and the Mankran-kesse with the lowest with a rating of 4. In descending order, the surface water potential in the sub-basin level is therefore as follows: Mankran-kuma, Biem, Dwinyan and Mankran-kesse-up. Even though the stream order of the Dwinyan River is 3, specific discharge data shows it has a higher surface water potential than the Mankran-kesse-Up which is of order 4. Looking at Table 2, three of the rivers are of the 4thorder with one (Dwinyan) being the 3rd order. The high variability of the specific discharges suggests that high stream-order rivers may not necessarily yield high specific discharges. Other hydrological or soil characteristics in the watershed may have a higher significant influence.
　　4.5 Estimation of Watershed Baseflow
　　It is a well-known fact that base flow contributes much of the streamflow [8, 9]. Thus, quantification of shallow ground water aquifers is important for sustainable ground water and surface water exploitation for irrigation and agro-ecosystem water supply.Previous research studies [9] estimated recharge to the shallow ground water aquifer from a base flow separation method considering evapotranspiration of deep rooted trees. The results of these studies [9] may be used for sustainable ground water development in a basin, not to exceed the recharge rate to the ground water aquifer. This is a case study that may be applicable to inland valley bottoms for rice production hence ensuring that the consumptive water use of rice does not exceed the aquifer recharge rates.
　　Baseflow at the watershed inlet and outlets were estimated using the recursive digital filter method as presented in the equation below [8, 10]:
　　(1)
　　where,
　　bk: Baseflow at time step k;
　　bk – 1: Baseflow at time step k – 1;
　　yk : Total streamflow at time step k;
　　BFImax: Baseflow Index (ratio of baseflow to the total flow);
　　a: Filter parameter.
　　In order to minimize the subjective influenceof filter techniques on choice of the BFImax, this study used a BFImax representative of the hydrological and hydrogeological characteristics of the watershed. Based on a preliminary analysis, this study used a BFImax of 0.25 for perennial streams with hard rock aquifers. Results from Fig. 6 indicate that at both the inlet and outlet, there is significant direct runoff with peaks corresponding to the high precipitation events as previously portrayed in Fig. 4. Results from the outlet which is the furthest downstream portion of the watershed indicate that there is higher discharge compared to the upstream inlet. The discharge at the watershed outlet would have been expected to be substantially higher but it was not the case. It is plausible that there could be water reserves trapped in the upstream sub-surface zones which take a while to percolate to downstream areas through interflow. The availability of this slow but gradual base flow contributes to sustained stream release of stored groundwater resources. The implication drawn from Fig. 6 is that there is sufficient subsurface seepage and flow to sustain inland valley rice production in both the upstream and downstream reaches of the watershed. This further supports results shown in Table 2 and Fig. 5.
　　 5. Conclusion
　　Results from this study indicate that at the sub-basin level, the Mankrankuma River has the highest surface water potential, followed by the Biem River, the Dwinyan River with the Mankrankesse-Up being the least. Therefore in terms of paddy field developments, the Mankrankuma and the Biem sub-basins would be highly suitable. The flow of the Dwinyan River in the Dwinyan sub-basin is insufficient and fluctuates highly with very high peaks and hence the need for storage provision such as reservoirs for paddy cultivation under semi-irrigation. The Kesse-up sub-basin also needs a storage system if a substantial area for rice production is to be effectively sustained. Despite the low flow for some portions of the watershed, the temporal patterns of the basin water resources indicate that annual stream flow in the watershed is substantial from May till end of October. As a result, thepaddy rice cultivation season should be focused in this window period with a possibility of harvesting in early November in order to avoid water stress. It is recommended that similar studies are conducted with more stream flow data in order to have a longer time-series representation of the overall water potential in the catchment. This should be supplemented with detailed water balance estimates to further evaluate both the surface and ground water resources in the watershed.
　　 References
　　[1] Statistics on World Hunger and Poverty Facts, FAO Report, Food and Agriculture Organization of the United Nations, 2010.
　　[2] M. Rosegrant, X. Cai, S. Cline, World Water and Food to 2025: Dealing with Scarcity, International Food Policy Research Institute, 2002.
　　[3] Agriculture in Ghana: Facts and Figures, Research and Information Directorate of the Ministry of Food and Agriculture, Ministry of Food and Agriculture, 2007.
　　[4] P.S. Thenkabail, C. Nolte, J.G. Lyon, Remote sensing and GIS modeling for selection of benchmark research area in the inland valley agroecosystems of west and central Africa, Photogrammetric Engineering and Remote Sensing, Africa Applications Special Issue 66 (2000) 755-768.
　　[5] T. Wakatsuki, E. Otoo, W.E.I. Andah, M.M. Buri, D. Kubota, Integrated Watershed Management of Inland Valleys in Ghana and West Africa, Eco-technology Approach: A Japan International Cooperation Agency/Crops Research Institute Joint Study Project Final Report, 2001.
　　[6] J. Oppong, M.M. Buri, T. Wakatsuki, Hydrological Characteristics and Water Quality of Rivers in the Mankran Watershed for Sawah production, Annual Report (2006-2007) for the Soil Research Institute and Kinki University Japan Joint Study Project, 2007.
　　[7] A.N. Strahler, Quantitative analysis of watershed geomorphology, Transactions of the American Geophysical Union 8 (1957) 913-920.
　　[8] K. Eckhardt, How to construct recursive digital filters for baseflow separation, Hydrological Processes 19 (2005) 507-515.
　　[9] H. Wittenberg, M. Sivapalan, Watershed groundwater balance estimation using streamflow recession analysis and baseflow separation, Journal of Hydrology 219 (1999) 20-33.
　　[10] B. Engel, K. Jae Lim, B. Thaller, S. Muthukrishnan, Y.S. Park, Web Based Hydrograph Analysis Tool (WHAT), Purdue University, USA, [Online], https://engineering.purdue.edu/~what/main.html.(Accessed on Mar. 2011)