Global science, local problems:

Seasonal climate forecast use in a Basotho village, southern Africa

Gina Ziervogel

Environmental Change Institute, University of Oxford, OX1 3TB, UK.

E-mail:

Prepared for presentation at the Open Meeting of the Global Environmental Change Research Community, Rio de Janeiro, October 6-8, 2001.

Abstract

Seasonal climate forecasts have been promoted as a means to lessen the problems of food insecurity in Africa. The realities of this are still to be seen. A focus on Lesotho, and in particular one village in southern Lesotho, highlights the information that these vulnerable small-scale farmers receive at present and how this aids their decision-making. An examination of the present use of traditional climate indicators, and access to and use of other agricultural information, provides insight into the role that seasonal forecasts could play in promoting sustainable livelihoods. Participatory research explores whether villagers think that seasonal climate forecasts could be of use to them, how they would use them and the present constraints that exist within the networks of local, district level and national communication and forecast dissemination. The case study is linked back to the bigger scale by highlighting the path that seasonal forecast development might need to take if it is to contribute to national and international development with regards to issues such as ameliorating food security and improving livelihood sustainability among vulnerable groups.

Keywords: seasonal climate forecasts, rural livelihoods, Lesotho

Introduction

Seasonal climate forecasts are often proposed as a means of factoring in climate variability in an attempt to ameliorate food insecurity in Africa (Broad and Agrawala 2000). Yet theories of food insecurity and livelihood insecurity emphasise the multiple dimensions that contribute to vulnerability with climate variability being only one dimension. Much of the work that assesses seasonal forecast applications has stated that certain conditions need to be met before marginal and other groups can utilise the forecast (Gibberd 1996; Orlove and Tosteson 1999; Stern and Easterling 1999). Broad and Agrawala (2000) stress the importance of the forecast being not only reliable but socially robust. If these conditions are met and it is acknowledged that climatic variation is only one part of a community or households’ vulnerability, what impact might seasonal climate have in supporting and accelerating the transition from vulnerable to sustainable livelihoods?

Predicting the future is never easy but over the last few decades, climatologists have improved on this with regards to the climate (Palmer and Anderson 1994; Murphy et al 2001). Seasonal climate forecasts are still a prediction, but they provide probabilities of rainfall totals and temperatures variations for the rainy season. Marginalised groups are often the ones who suffer most in highly variable climates, as they often do not have the resilience either to prepare for or to recover from the shocks that climate extremes tend to bring. Most of these groups have local forecasting knowledge that may become threatened or less reliable due to climatic changes (Roncoli et al. 2000). Translating the forecast into a means for harnessing climate variability and subsequently into reducing food insecurity is a complex pathway. The best way to understand it is to examine the local realities and then work up to what the implications might be on a global scale (Wilbanks and Kates 1999) .

How could vulnerable households in southern Africa use seasonal forecasts? In sub-Saharan Africa 20% of the Gross Domestic Product (GDP) and more than two-thirds of the labour force is accounted for by agriculture (World Bank 1998). This translates into food security in a number of ways as outlined by Devereux and Maxwell (Devereux and Maxwell 2001). Firstly there is the food that is produced by rain-fed agriculture but it also translates into livelihoods, a market, raw materials, foreign exchange and surplus as the trickle down from agriculture. It seems reasonable to expect that a forecast of how much rain there might be would be of direct relevance to managing agriculture and other climate-related activities by maximising the use of available resources and minimising the impact of extreme situations (Stern and Easterling 1999). For example crops suited to the predicted rainfall could be planted. If a drought was forecast sorghum would be a sensible crop to plant as it requires less water than maize. Forecasts could allow for appropriate water management regimes. If heavy rains are forecast then it might be sensible to release water instead of backing it up in dams and if below normal rains are forecast then water could be used more conservatively. The forecast could be used to gauge potential markets and trends. ‘Good’ rains might result in an excess of grain for sale, so it might be wise to plant another crop that would have greater market demand.

An assessment of how the forecast might benefit marginal groups can best be reviewed by looking at how the forecast is presently being developed, disseminated and used in particular cases (Eakin 2000; Mukhala 2000; Roncoli 2000; Vogel, 2000; Phillips 2001). This paper uses a case study of a rural village in Lesotho to highlight the complexities associated with providing the forecast to marginal groups in rural areas in developing countries and how it might be integrated into rural livelihoods. Research based on participatory fieldwork provides explanations for some of these questions at the local scale and it is expected that these results can be drawn upon for exploring the role that seasonal forecasts might have in other developing regions. Lesotho was chosen as a country that is small enough to look at the national operations as well as having many marginal groups that are highly dependent on rain-fed agriculture (Chakela 1999). There is also a lot of work being done in Lesotho on livelihood security, which complements this research (Gay and Hall 2000; CARE 2001; Turner 2001).

Forecasts in southern Africa

A seasonal forecast is the prediction of what the chance of the total amount of rainfall being above, below or within the normal range is for the coming three to six months. It is based on the principle that the ocean and atmosphere are highly dependent, with sea surface temperatures providing an indication of future atmospheric perturbations (Washington and Downing 1999). Ocean temperature anomalies can persist for months or years, driving the flux of heat to the atmosphere, which affects global temperature, rainfall and circulation patterns (Mason et al. 1996).

The forecast is not deterministic but probabilistic and works on the chance of monthly rainfall totals falling into a specified category. It is common for terciles to be used, which means that rainfall is predicted to be either below normal, normal or above normal (as seen in Figure 1). Seasonal forecasts are based on one of two methods. The numerical approach uses General Circulation Models (GCMs), integrated with sea surface temperatures to give a prediction of how much rain there will be for up to 50 days ahead (Washington and Downing 1999). This method is used by well-established groups such as the United Kingdom Meteorology Office (UKMO) and the European Centre for Medium-Range Weather Forecasts (ECMWF) as it requires a lot of computing power for the models to be initialised and run with different starting conditions, together known as an ensemble. The other method uses statistical analyses of past rainfall data. Historical rainfall records are statistically analysed for trends and multi-annual variability that can account for 30% of the rainfall variability (Washington and Downing 1999).


Figure 1: Forecast for total rainfall in October, November and December 2000 in Lesotho. There is a 45% probability of there being below normal (B) rainfall in Region 1 and a 45% probability of there being near normal (N) rainfall in region 2.

The rainfall season in southern Africa extends from October to March (Washington and Downing 1999). In the second half of the season (January-March) the rainfall originates in the tropics, which is thought to be more predictable than the predictions of rainfall that originates in the mid-latitudes during the first half of the season (Harrison 1984; Tyson 1986; Mason et al. 1996) . In September the annual SARCOF (Southern African Regional Climate Outlook Forum) is held. This provides an opportunity for regional collaboration by providing a forecast for the entire southern Africa region. Meteorologists and users from the SADC (Southern African Development Community) countries gather to develop a consensus seasonal forecast for the region and users comment on how they have used the forecast and the improvements they would like to see (O'Brien et al. 2000).

There are many constraints to forecast use and dissemination. Two aspects that are dependent on improved skill from the meteorologists’ side are the spatial and temporal limitations. The spatial scale is often not good enough if the forecast is to be used on a local scale, for example at the village level. There are usually only a handful of regions that the country is broken down into with different predictions for each one but not information on local climatic variation. The temporal scale can also be problematic as it gives the prediction for the total rainfall over three months and does not usually say when during that time the rainfall is likely to occur or when the rains might start.

Case Study: Lesotho

In 1820, during the Lifaqane, Moshoeshoe and his Basotho (collective name for the inhabitants of Lesotho) followers were forced across the Caledon River, from what is now the eastern Free State, into the foothills and mountains of Lesotho (Ministry of Natural Resources 2000). Before these groups arrived there was evidence of small groups of San, who were hunters and gatherers and did not place extreme pressure on the land (Chakela 1999). Within a hundred years nearly all the wild animals and sweet grass had gone due to the rapid expansion of livestock and the arrival of the plough. The first census, in 1875, counted 128 000 people and by the 1996 census there were 2 100 000 people in Lesotho (Bureau of Statistics 1996). One of the results of this increase in land pressure has been the severe degradation that now plagues most agriculturally viable land, which is estimated to be 11% (Chakela 1999).

The Lesotho environment is particularly harsh, both because of its terrain and climate. The altitude ranges from 1388 – 3 482m above sea level. The country that is totally landlocked by South Africa, covers an area of 30 000 square kilometres (Ministry of Natural Resources 2000). The mean annual rainfall ranges from 426mm in the lowlands to 1 097m in the highlands, with high variations in the different regions (Hyden and Sekoli 2000). The country is divided into four agro-ecological zones, with the lowlands in the western portion of the country rising to the foothills, and then to the highlands in the east. The highlands account for 60% of the land cover and have the lowest concentration of people. In the southern part of the country there is an area known as the Senqu River Valley that has different agro-ecological characteristics and is often classified as a fourth zone. This research was undertaken in the southern part of the country, where all four agro-ecological zones are encountered.

Agriculture continues to support the majority of the population in Lesotho despite the myriad challenges faced (Gay and Hall 2000). Farming occurs mainly on marginal land that is often steep, eroded, unfertile and to make matters worse has a highly variable climate (Ministry of Natural Resources 2000). Although there has been an increase in urbanisation, there remains a large proportion of the population that are dependent on agriculture despite the declining quality of the soil and limited returns (Turner 2001). People are trying to survive and agriculture provides a means of doing so as everyone is entitled to land by law (although it is limited); labour is readily available and seed is usually kept from the previous year’s harvest.

Although agriculture plays an important role in the lives of Basotho, some have said that they cannot be called farmers because very few households engage solely in agriculture. Wage earning is the most favoured livelihood strategy and if that is not possible there are many other strategies that households might engage in (Turner 2001). These include the sale of firewood, making bricks, sewing, selling local beer (joala), piece jobs in South Africa or elsewhere to name a few (Gay and Hall 2000). In Livelihoods in Lesotho (Turner 2001), the Basotho were asked to compile their own profiles of wellbeing. To assess wellbeing they considered what a household has access to in terms of livestock, fields, assets such as housing and farming equipment, access to wage income, children in school, food security, ownership of a business, disabled members and whether the household can hire workers itself (Turner 2001). Some of these strategies and resources are directly impacted on by the climate and others might be affected down the line but it is clear that the climate has an impact on all livelihoods.

The state of seasonal forecasts in Lesotho

Lesotho Meteorological Service (LMS), housed in the Ministry of Natural Resources, are responsible for the development and dissemination of the seasonal forecast. They have produced forecasts since 1997 and independently since 1998[1]. At present the forecast is announced at a National Workshop at the beginning of October. The rains start around October, so anyone who wants to make preparations in the previous months would not have heard the forecast in time. Public and private representation at the workshop is also a problem. People are sent letters inviting them to the workshop and there is an announcement in the newspaper that invites all interested parties. Last year the letter went out on the Tuesday and Wednesday when the meeting was held that Friday. There were about forty people at the meeting including about four commercial farmers and the rest of the people were from media or government.