Implication of climate change and variability on stream flow in Iringa region, Tanzania

Agriculture remains the mainstay of the economy of many sub-Saharan Africans and central to the survival of millions of people (FAO 2004). It provides employment to an average of 65% and contributes to about 32% of the GDP (gross domestic product) (AASR 2013). In Tanzania, the sector contributes to about 50% of GDP, 80% of employment, and 75% of export (AASR 2013). Despite being crucial to the economy of the continent and the country in particular, it is among the main sectors affected by climate change and variability (CC&V) and its associated impacts (Kandji et al. 2006; Mary & Majule 2009). The negative effects of CC&V on agriculture include unreliable rainfall pattern, the decline in rainfall amount, drought, floods, and outbreak of crop pests and diseases (URT 2012). To reduce vulnerability levels, indigenous farmers have adopted various coping strategies aimed at improving storage and utilization of water resources. These include various techniques of water harvesting and utilization in sustaining crop yields (Majule & Mwalyosi 2005; Kassian 2012). The same techniques of traditional irrigation conducted by using shallow canals are practiced in the northern state of Himachal Pradesh (Gover et al. 2013).

Those techniques have different names in different localities, including, for example, kuhls in the northern state of Himachal Pradesh, India (Gover et al. 2013) and ngoro or matengo pit in Ruvuma region (Sims et al. 1999; Yanda et al. 2005), majaluba in Mara region (FAO 2001; Yanda et al. 2005), ndiwa in west Usambala mountains (Sokoni & Sechambo 2005) and ‘vinyungu’ in Iringa and Njombe regions (Majule & Mwalyosi 2005) practiced in Tanzania.

The vinyungu system is practiced in areas that are moist throughout the year; sometimes the system depends on water from rivers for irrigation (Majule & Mwalyosi 2005; Munishi et al. 2011). As a result, vinyungu are mostly located along rivers, around water catchments and/or within wetlands. However, even vinyungu systems tend to be affected by CC&V through the reduction of water level in rivers, which most farmers rely on for irrigation (Majule & Mwalyosi 2005; Magembe 2007). In addition, it tends to cause drying up of wetlands suitable for vinyungu farming. In most cases vinyungu is associated with low water use efficiency including diversion of water from river channels for irrigation and some practices that lead to the collapse of riverbanks and drying up of wetlands and downstreams (Mallango 2001). This results in the collapse of riverbanks and drying up of wetlands and downstreams. It is mainly due to this that acts and by-laws to govern the conservation, management, and development of water sources, riverbanks, and beaches or coasts have been passed (Madulu 1998; Mkavidanda & Kaswamila 2000; Kassian 2012; Kisanga 2013). Despite the fact that most farmers are aware of the acts and by-laws, they do not comply with them. Consequently, governmental and non-governmental institutions in Tanzania had either to improve or limit vinyungu farming.

Even though there are various studies on vinyungu practice and its socio-economic importance (Mkavidanda & Kaswamila 2000; Mallango 2001; Majule & Mwalyosi 2003, 2005; Magembe 2007; Kangalawe et al. 2011; Munishi et al. 2011), little has been done regarding the implication of CC&V on the Lyandembera River flow used for vinyungu irrigation for food security and income generation in the study area. Therefore, the current study aimed at filling such a knowledge gap in Iringa Rural District by assessing the implication of CC&V on the river flow of Iringa Rural District in the southern highlands of Tanzania. The relationship between rainfall and river flow that is used for traditional irrigation and various adaptation strategies for water shortage used for irrigation were investigated. The results are envisaged to stimulate further research on proper adaptation strategies for water shortage and improvements in crop production, income generation, and river flow.

The study was carried out in Iringa Rural District, Iringa region located in the southern highlands of Tanzania. The district is among four districts of the region lying between latitude −7 °33′50.36 and longitude 34 °52′53.73, practicing the vinyungu system using water from rivers for irrigation. The district has two ecological zones: the midland zone lying between 1,200 and 1,600 m above sea level (ASL) with rainfall ranging between 600 and 1,200 mm per year; and the lowland zone ranging between 900 and 1,200 m ASL receiving rainfall between 500 and 600 mm per year.

Both primary and secondary data were used in the study. Primary data collection was undertaken using household questionnaire, key informants’ interview, focus group discussion (FDG), and direct field observation. Several research tools were used to acquire local perceptions on the trend of rainfall and river flow, causes, indicators for the current situation in Lyandembera River, and by-laws concerning human activities along the river. One hundred and eighty-nine households, forming a sample size of 5% from each village, were surveyed during household questionnaire as recommended by Boyd et al. (1985). Heads of household or other members were selected strategically with the help of the Village Executive Officer (VEO). Respondents’ selection was based on those aged 30 and above and practicing the vinyungu farming system along the Lyandembera River. The VEO, District Agricultural Officers and Agricultural Extension Officers were among the key informants to provide information, for example, village by-laws; a FGD of eight members in each village was conducted as recommended by Krueger & Casey (2000). Each group involved both males and females aged 30 years and above. This age selection was based on experience of rainfall for about 30 years and practicing vinyungu farming along the Lyandembera River. Key informant interviews and FGD were conducted with the aid of a checklist. Guided and unguided direct field observations were used to cross check information including information about vinyungu practices.

Secondary data on the farming system were obtained through a literature review. Climatic and river flow data were obtained from Iringa Rural District and Rufiji Basin Water Offices (RBWO) based in Iringa municipality. Monthly rainfall and Lyandembera River flow data from 1997 to 2014 were collected at Kalenga and Ilongo stations, respectively. These data were used to investigate whether there is a relationship between the trend of rainfall and river flow in Lyandembera River.

Qualitative data were produced in an outline form to perform content analysis after organizing and coding into categories. Quantitative data were analyzed using Statistical Package for Social Sciences (SPSS) program (version 20). Cross tabulation was used to compare different parameters from the two villages during data analysis. Then, information from SPSS was presented by graphs using Microsoft Excel software. Furthermore, the annual trend of rainfall and river flow data were analyzed by using Mann–Kendall's statistical test and linear trend analysis, and presented graphically using Microsoft Excel.

The decline of river flow has negative impacts on crop production, especially on the vinyungu system in the study area as reported by 69 and 87% of the respondents in Lumuli and Ifunda villages, respectively, in the household survey. This could also have an impact on crop production in other areas that depend on water from the river for irrigation. It has been projected that key irrigation areas in Tanzania will face a decrease in rainfall as well as an increase in temperature that will increase evaporation of surface water resulting in water shortages for irrigation (TACRP 2014). Furthermore, URT (2012) showed that reduced amount and changed rainfall seasons have a negative implication on some perennial rivers such as the Great Ruaha in Iringa region.

These findings are parallel with local people's perceptions revealed during the household questionnaire. It was reported that amount of rainfall, onset and end of the rainfall period in the study villages have changed compared to the past 30 years. In Lumuli village, 56% of the respondents stated that in the past 30 years rainfall used to start in November and end in May. This is contrary to recent years where it starts in December and ends in April. Likewise, in Ifunda village, 51% respondents reported that rainfall in the past used to start in November through to May. However, in recent years, 58% of the respondents stated that it starts in November and ends in April. This has been directly observed in both villages during data collection done in April, a period for nyahengi, a local name of rains that is a sign of the end of the rainfall period. A study by Kassian (2012), in Mufindi District, Iringa region indicates the same type of rainfall (nyahengi) is a type of rainfall which used to be present in May up to July after the end of heavy rains. The rain was very crucial to the crops that take a long time to mature, planted in the middle of the rainy season and late planted crops (Kassian 2012).

Findings from this study revealed that there is a tremendous decline in river flow used for vinyungu irrigation, resulting in a decline in crop productivity in the area. To cope with and maintain crop productivity, various practices have been employed and their costs were investigated.

River diversion was reported by one of the key informants to cause water shortage to the downstream farmers in Lumuli village. In addition, fetching water direct from the river has been reported to cause destruction of natural vegetation and hence soil erosion in riverbanks. This could be exemplified by the visibility of stones in Ifunda village which local people associate with the decline of river flow (Figure 5(b)).

Reduced size and drying of wetlands was the second problem mentioned in the household questionnaire, in Lumuli village by 37% and 32% of the respondents in Ifunda. This also has been revealed in land-use/cover change analysis from satellite images in both wards. Images showed a decline of wetlands from 57.673 km² to 42.445 km² in Ifunda village and from 51.679 km² to 39.445 km² in Lumuli village from 1984 to 2014 (Table 1). This may be due to the fact that Lumuli village is located near wetlands surrounding the village. The vinyungu farming system in the village has a direct impact on wetlands by reducing them at the expense of expansion of the farming system. According to Magembe (2007) and Munishi et al. (2011), availability of moisture throughout the year in the changing climate forces farmers to expand cultivated areas.

River sedimentation and vinyungu fields flooding were among the problems reported by a minority (5%) of the respondents in both study villages. These problems were reported to occur during heavy rainfall periods, which are accompanied by the destruction of vinyungu. Villagers in Lumuli reported that the problem was more pronounced in Lumuli village probably because of farming practices in wetlands. Wetlands act as a sponge to store water. This might further be due to the location of vinyungu in the lower stream in Lumuli village relative to Ifunda village. As a result, some of the farmers in Lumuli village abandoned the vinyungu system. Abandoned fields have positive effects on the environment as they increase wetlands. Similarly, Kangalawe et al. (2011) and Kassian (2012) have reported these problems to be prevalent in Mafinga District, Iringa region.

Generally, irrigation methods used have implications for the environment, as the practice may result in competition for water usage for different purposes such as domestic use. It may also result in the destruction of other living things such as wetland vegetation that depends on water for its survival.

There are some efforts which are in place at the village level to overcome the environmental impacts of the vinyungu system in the study villages. These include restriction of cultivation near water sources and within 60 m from valley bottoms and river sources (URT 2004) and cutting water-conserving trees such as S. cordatum locally known as mivengi. This was reported during the FGD in both villages. A study conducted by Kassian (2012) in Mufindi District showed that trees such as S. cordatum, ‘misombesombe’, ‘mivelevele’, ‘misusuriana’ and ‘midzobwe’ are used as water-conserving trees and some of them are planted in wetland areas. However, in Kilolo District, Kyando (2007) reported that seminars facilitated by the World Wide Fund for conservation of nature (WWF) in association with the staff from RBWO are conducted to train farmers in proper ways of conserving the environment including valley bottoms.

Scientific evidence from this study is aligned with local perceptions that climate change in terms of rainfall decrease has implications on the trend of river flow. The trend of the river flow corresponds to the trend of rainfall which shows a slight declining pattern in the last 18 years. This has an impact on crop production in vinyungu fields located along the Lyandembera River, which depend on water from the river for irrigation. In maintaining crop productivity and soil moisture, various methods reported to be employed include increasing the depth of water wells in the fields, diverting the river through water channels to the fields, and increasing the use of water from the river. These affect river flow. Therefore, this study recommends that central and local government should put more effort into enforcing laws and by-laws that restrict any human activities such as agriculture to be conducted on wetland and within 60 m from the river and around water sources. Furthermore, this paper recommends that more research should be conducted on proper water management and irrigation methods for sustainable river flow for both agriculture production and environmental management for different farmer typologies.

This research is supported by funding from the Department for International Development (DfID) under the Climate Impact Research Capacity and Leadership Enhancement (CIRCLE) programme.