In this chapter, we review the main conclusions drawn from our in-depth case study of the 1991/92 Southern African drought. We then discuss many of the lessons that have been drawn from that experience. Since the 1991/92 "Famine Scare" in Southern Africa, numerous changes have been made by ENSO forecasters and members of the food security community that have enhanced the potential utility of ENSO information in the SADC region. There remain, however, obstacles to the effective routine use of ENSO information including forecasts, for food security in Southern Africa.

7.1. Conclusions

Importance of Historical Context

The 1991/92 drought in Southern Africa occurred in a specific historical context, one which left the SADC region particularly vulnerable to the effects of drought. The region suffered from poor rainfall and crop production in the 1990/91 growing season thus grain stocks were already precariously low in March 1991. In addition, the area devoted to food crop production was on the decline because of government incentives that made the production of cash crops such as tobacco and cotton more attractive. For example, large-scale producers in Zimbabwe diversified their crop portfolios away from maize production. This left communal farmers to supply the country (and much of the SADC region) with maize. Unfortunately, many of these farmers lacked the advanced technologies needed to increase their output and were often farming marginal (e.g., less productive) lands. Low food stocks as well as reduced production placed the countries of Southern Africa in a vulnerable position as they entered the 1991/92 growing season, even before they were faced with the drought.

This vulnerability was further increased by the introduction of ESAPs and other economic reforms in several SADC countries, apparently with litle or no regard for climate variability's impacts on agricultural production. In their efforts to reduce their government's expenditures, countries such as Zimbabwe and Zambia reduced food subsidies. This caused food prices to increase sharply. As a result, a larger number of people living in those countries found themselves dependent on food aid, even before the impacts of the 1991/92 drought were felt. Also, in the late 1980s and early 1990s Zimbabwe tried to reduce its expenditures for the storage of food reserves and exported large amounts of grain. During the 1991/92 growing season (at the same time there were reports about pending food shortages), Zimbabwe exported 35 percent of its 656,000 tonne opening stock. By the time it became absolutely clear that Zimbabwe's 1991/92 maize production would be drastically reduced due to the drought, there was not enough maize in reserve for the government to provide any significant relief to its own people.

The 1991/92 drought introduced forces that conflicted directly with the objectives of the structural adjustment programs. At a time when governments were supposed to be cutting expenditures in order to control budget deficits, they were actually forced to increase expenditures in order to avert famine. For example, in Zimbabwe, the government increased prices paid to maize producers in order to encourage greater production in the 1992/93 growing season. Maize imports which became necessary because reserves had been drastically reduced, cost as much as four times the subsidized price at which the government sold them to industrial millers. The budget had to absorb these increased expenditures at a cost to the government of Zimbabwe of approximately US$300 million, or 11 percent of its GDP for 1992/93.

The SADC regional food security community issued warnings of an impending food security crisis as early as March 1991. These warnings were issued throughout the year, even before information about the possibility or likelihood of an ENSO event and the related drought became available in the region.

ENSO Information in the SADC Region

At least one ENSO forecast model, the Cane-Zebiak model, was beginning to predict the onset of a warm event by the end of 1991 as early as January of that year. By April 1991, the two main models used by the U.S. Climate Analysis Center (one of the major sources of ENSO information for the SADC region in 1991) both suggested an ENSO event would soon be developing. Unfortunately, these models were surrounded by a great deal of "noise." For example, the oceanic and atmospheric indices being monitored were far from conclusive and the eruption of Mt. Pinotubo in August 1991 distorted satellite readings of sea surface temperatures. It was not until September 1991 that model output and actual observations enabled the international ENSO forecasting community to agree an ENSO event was in its early stages. The first mention of the rainfall deficiencies that accompany ENSO events came in November.

Although ENSO information, including a forecast, was "out there" throughout most of 1991, few decision makers within the SADC region had access to this information. At the time of the drought, there was no formal structure or process in place for disseminating ENSO information. Few organizations within the region received ENSO information and none sought it. It is unclear what, if anything, they did with the information once it had been received.

Some organizations, such as the Drought Monitoring Centre in Harare, did receive ENSO information and distributed it to government officials in the SADC region. However, agricultural and food security decision makers were for the most part unfamiliar with ENSO and its link to drought in Southern Africa. As a result, ENSO-related information often sat on bureaucrats' desks while they attended to "more pressing" economic and political matters. In general, we find ENSO information did not play a significant role in the national, regional, and international responses to the 1991/92 Southern African drought.

Politics and ENSO Information

Early in 1991, Zimbabwe's Grain Marketing Board (GMB) acknowledged there would be a gap in maize supplies from the end of 1991 until the harvest of the 1991/92 crop in early 1992. The Board budgeted for imports to fill this gap at the beginning of the new marketing year (1 April 1991), and sought the foreign exchange it would need to purchase between 100,000 and 200,000 tonnes of maize. It was assumed the maize could be obtained from the Republic of South Africa.

In other parts of the Zimbabwe government, however, officials were skeptical about the need to import maize. Some felt that rural farmers were storing extra maize on their farms with the hope that the fixed producer prices would be increased by the government. In addition, Zimbabwe is traditionally a maize exporter; thus, importing maize was contrary to standard operating procedure. Due at least in part to this skepticism as well as ESAP-related conditions, the foreign exchange necessary to finalize the purchase of imported maize was not forthcoming. This remained a problem even after officials with the SADC REWU informed the government of Zimbabwe that the region was likely to experience drought-related food shortages in 1992. Many Zimbabwean government officials did not take this warning seriously and assumed the drought would not be region-wide. They also assumed that if needed, reserves could always be acquired from South Africa.

The situation in Zimbabwe is not unique; it illustrates the political nature of drought and food insecurity in many parts of the world. As Glantz (1996a) notes, "No government likes to admit that it cannot feed its people. Thus, early warnings about impending food shortages and the food needs assessments that follow are often politically charged statements. In fact, governments frequently delay announcements of impending food shortages, hoping that they will not develop into full-blown crises."

Drought as an Event

In general, we found that national, regional and international responses to the drought were initiated when the rains had failed, crop damage had become visible, and evidence was presented that the rains were unlikely to return in time to produce a reasonable harvest. In other words, most officials at various levels of government as well as humanitarian aid donors tended to view the 1991/92 food shortage as a specific discrete "event" rather than as a process that had begun in early 1991. This perception influenced the type of information decision makers were either seeking or were receptive to. This represents an example of selective inattention or not paying attention to information they deemed unreliable or unimportant. Specifically, they were waiting for "hard evidence" that a drought had indeed adversely affected the 1991/92 harvest and would result in drought-related food shortages. ENSO information for those aware of this Pacific Ocean phenomenon and its teleconnections around the globe which suggested that a drought was likely to develop in the region, was not the type of "hard" evidence decision makers were apparently looking for.

That drought was being viewed as an event rather than as a process in 1991/92 can also be inferred from the nature of the early warning systems in the region at the time. In 1991/92, "early" warning for food security meant providing information on food supply and demand so that governments, donors, and relief agencies could make plans for the provision of humanitarian and other types of interventions as soon as possible. Warnings were considered "early" in relation to the robustness of the harvest near the end of the growing season rather than to the planting of crops at the beginning of the season. Again, early warnings were largely based on hard evidence of actual food deficits and the appearance of the adverse impacts of drought on crops rather than on the increasing possibility that drought-related food deficits might occur. Glantz has noted that drought alone seldom leads to severe food shortage. Drought and some other factor(s) combine to generate severe shortages.

Would Earlier ENSO Information in 1991/92 Have Made a Difference for Food Security?

Based on information obtained from interviews with individuals involved in the 1991/92 drought in Southern Africa, a number of actions have been identified that, in theory, could have been taken to mitigate the impacts of the drought on society. For example, countries in the region could have placed orders for grain imports earlier in 1991 and, perhaps, have obtained food supplies at a cheaper price. An earlier response would also have enabled governments to take advantage of cheaper transportation alternatives and avoid expensive transport bottlenecks. Likewise, relief agencies could have mobilized food supplies earlier and better coordinated food shipments. Zimbabwe could have reduced or halted its maize export program much earlier, using those supplies to increase its domestic reserves, saving approximately US$41 million. Farmers in the SADC region could have sold their livestock at higher prices, which would have enabled them to purchase needed food supplies, when the drought reduced the 1991/92 harvest. In addition, farmers might have been able to plant more drought-resistant crops such as millet and cassava instead of maize.

Despite these potential opportunities, there were numerous social, political and economic constraints that would have inhibited the use of earlier ENSO information during the 1991/92 drought situation. Politically, decision makers are often most concerned with short term benefits. In other words, political incentives to focus on the short term, while drought response and food security require decision makers to take a relatively long term perspective. Perhaps more important in this situation most government decision makers were at the time either unaware of the connection between ENSO and Southern African drought and they did not believe the information they may have received to be either credible or reliable. As a result, bureaucrats in the ministries of agriculture within various SADC countries often failed to pass ENSO-related information along. Agricultural extension officers, who would have been the primary link between these ministries and farmers, were also untrained in how to use ENSO information in at their decision making level.

Had farmers received ENSO information earlier in 1991, they would not necessarily have been able to take full advantage of that information. For example, in order to switch to drought resistant crops, farmers would have needed to obtain the necessary inputs (such as seeds) in a timely fashion. Many individual farmers may not have had the resources available to obtain these inputs. Likewise, many of the SADC governments would not have been in a position to assist farmers financially in making the switch.

As for the ability of humanitarian relief organizations to respond to the earlier availability of ENSO information, this too would have been unlikely in 1991/92. Economic resources of these agencies are generally stretched very thin. Because they are unable to commit funds very far in advance they must often rely on signs of an actualsevere drought situation rather than indicators of a potentialbut less than visible situation.

Based on our assessment of the 1991/92 Southern African drought situation, we conclude it is unlikely that earlier ENSO information would have made a great deal of difference in attempts to mitigate the impacts of the 1991/92 drought.This is largely because of the fact that there was no organized mechanism for widely distributing the information. And, even if decision makers had received such information in a more timely manner, they were unfamiliar with the link between ENSO and drought in the SADC region and how to use ENSO information to make food security-related decisions. In addition, there were numerous social, political, and economic constraints on the use of the information.

7.2. Lessons Learned

As a result of the 1991/92 drought and near catastrophic famine in the SADC region, changes have been made in the ENSO forecasting system and in food security policy at both the regional and national levels in Southern Africa. Most importantly, there is also a greater awareness in the region of the strentgh of the teleconnection between ENSO in the Pacific Ocean and drought in Southern Africa.

Changes in ENSO Forecasting in the SADC Region

Interest in ENSO forecasting in the SADC region has come a long way in just the few years since early 1992, particularly with the addition of South Africa as a member of SADC. In 1991/92 regional forecasters relied heavily on analyses provided by the U.S. CAC (now CPC) and, to a lesser degree, the Australian BOM. Within the region, there were few individuals or groups monitoring ENSO indicators on a regular basis and conducting their own analyses.

Today, SADC's forecasting ability has increased with the addition of South Africa's expertise. The primary ENSO forecasters include, but are not limited to, the South African Weather Bureau (Pretoria), the Climatology Research Group at the University of Witwatersrand (Johannesburg) and Dr. Mark Jury at the University of Cape Town. Researchers at the University of Pretoria also issue forecasts on an ad hoc basis. These groups have formed the South Africa Long-Lead Forecast Forum in an effort to coordinate their forecasts and improve collaboration.

The Drought Monitoring Centre (DMC) in Harare has also increased its capacity to monitor ENSO. In 1991, the DMC included ENSO information received from the U.S. CAC in its monthly bulletins. Today, the DMC has direct contact with the Africa desk at the CAC and monitors ENSO as well as other oceanic indicators, such as Atlantic and Indian Ocean SST anomalies, on a regular basis. The DMC also has the capacity to conduct its own analyses. In September 1995, the DMC coordinated a seasonal forecasting meeting during which it provided a forecast using ENSO information of the upcoming growing season. The meeting was initiated by the Zimbabwean sugar industry, with participants attending from a number of SADC countries.

ENSO information from the different members of the regional forecast community is distributed to a variety of actors. The DMC has a mailing list of approximately 100 individuals, including the national weather services in each SADC member country, relief agencies, hydrologists, bankers and investors. The Climatology Research Group at Witwatersrand University faxes its monthly seasonal rainfall outlook (which it began producing in 1992) to the meteorological services and National Early Warning Units in each SADC state as well as to nearly 100 commercial users. In addition, each of the National Early Warning Units receives ENSO Advisories from the U.S. Climate Prediction Center.

ENSO forecasters in the SADC region are increasingly making efforts to develop linkages with their users in order to better understand the needs of the users. In the case of the Climatology Research Group, these interactions are generally initiated by individuals and organizations who contact the group in search of additional information and/or clarification. The South African Weather Bureau recently conducted a "Workshop on the Use and Benefits of Seasonal Outlooks." The purpose of the workshop was to educate users on the utility of long-lead forecasts and to assess the needs of end users. SADC/UN FAO also recently met with forecast users to discuss problems of getting climate information to the right people.

In October 1996, the Office of Global Programs of the U.S. National Oceanic and Atmospheric Administration convened a workshop in Victoria Falls, Zimbabwe on reducing climate-related vulnerability in Southern Africa. The workshop brought together producers and users of climate forecast information in the SADC region. The workshop was designed to demonstrate existing predictive capability in the region as related to identified needs in the agriculture, food security, water management, and health sectors. It was also an opportunity to identify requirements for improving forecast production as well as additional applications in other sectors. This is yet another example of efforts in the SADC region to improve communication between ENSO forecasters and users so as to improve the potential utility of ENSO information.

Changes in Regional Food Security Policy

Until well into the 1991/92 growing season the regional food security community did not incorporate ENSO information into its drought planning or early warning efforts for a variety of reasons. The reliability of climate forecasts were generally viewed with skepticism and the national early warning units were not set up to make predictions about the upcoming season. Rather, their mission was to report on the current state of crops and rains. However, today, there is a greater emphasis within FSTAU and its programs on providing predictions of climate outlooks for upcoming growing seasons. In addition, the DMC is using ENSO information as well as other indices (such as pressure systems in the Indian Ocean) to predict rainfall potentials in Southern Africa.

Changes in Awareness of the ENSO-Drought Linkage

The majority of individuals contacted in the past year (1996) in the SADC region believes few people in areas related to food security had heard of El Niño or ENSO by 1991/92, and even fewer were aware of its connection to regional drought. This lack of awareness of the event and its possible teleconnection to Southern Africa was one of the major obstacles to using an earlier ENSO forecast then. If people are not familiar with the phenomenon or its impacts, they are unlikely to pay attention to forecast information. Several of those individuals in the food security community in Southern Africa who were familiar with ENSO remarked that the link with regional drought was perceived to be tenuous at that time and that research was in its infancy and thus not viewed as highly credible. In other words, they had little confidence in the information provided and would not have been likely to put much weight on the use of an earlier forecast. In fact the level of awareness of ENSO's possible effects on Southern Africa among international humanitarian aid agencies and organizations was very low to non-existent.

Today evidence indicates that a growing number of policy makers in the region are aware that ENSO is related to drought in Southern Africa (UN FAO, 1996). Virtually everyone in the food security community that was interviewed acknowledged the importance of ENSO information in making food security decisions. This knowledge has apparently filtered down to the level of district farmers who, according to one grain trader, "talk about El Niño in the pubs." In addition, individuals and organizations outside the food security sector are increasingly seeking ENSO information to help them in their business planning. For example, the sugar industry in Zimbabwe contacted the DMC seeking information on ENSO and the 1995-1996 growing season. In South Africa, the state electric company supports long-term climate forecasting research. Several researchers reported having been contacted for ENSO information by commodity brokers dealing in sugar and tobacco and by financial institutions.

7.3. The Utility of ENSO Information for Food Security in 1996

The improvements in our understanding of ENSO and in ENSO forecasting, in food security policy and in awareness of the ENSO-Southern African drought link since 1992 have certainly improved the utility of ENSO information for regional food security. More organizations in the SADC region are aware of or are monitoring ENSO indices and disseminating information to a wider audience than was the case in 1991/92. Food security decision makers have integrated the need to consider drought in their planning process and have in many cases developed lines of communication with the forecasting community in order to stay informed about the onset of an ENSO event. Increased awareness of ENSO and its connection to regional drought increases the likelihood that individuals and organizations receiving ENSO information will direct at least some attention to the problem, rather than putting the information in a pile of papers on a bureaucrat's desk for eventual perusal.

This is not to say, however, that should an ENSO forecast be issued today that the response necessarily would be more timely or efficient than it was in 1991/92. Potential users are still faced with numerous obstacles to the use of ENSO information in decision making. These include questions about the reliability of forecasts, delays in the regional availability of forecasts, a lack of understanding of what the forecasts mean, as well as concerns about the utility of regional scale forecasts for local-level decision making.

7.4. Continuing Needs of ENSO Information Users

In feedback received from various individuals contacted during this study, several changes in the way that ENSO information is currently presented were suggested. Such changes might improve the ability of decision makers to use ENSO information, including forecasts, in their decision making process. Members of one research organization indicated they would be more likely to use ENSO forecasts in their work, if the forecasts were accompanied by an understandableindex of reliability. They felt their names (and jobs) would be on the line, if they were to make faulty decisions based on a forecast that proved to be erroneous. They noted that they would have greater confidence, if the forecasts included some measure of skill (high skill at that).

A reliability index might also help users clear up or at least understand some of the "noise" that usually accompanies ENSO information. Users often receive information from several sources and it can be difficult for them to evaluate which information from what source should be given greater weight. One group suggested that it might be useful to have an independent evaluator review various forecasts and assess their reliability before the forecasts reach users in the SADC region.

Several respondents noted the spatial scale of ENSO forecasts is often inappropriate for the levels of decision making in which they are involved. A regional forecast, for example, does little to help a small-scale farmer living in a particular agricultural district to make planting decisions.

It was also suggested that farmers are more interested in knowing the distribution of rain throughout the season rather than the overall rainfall characteristic of a season. For example, a small amount of rainfall during the critical flowering stage of the maize crop can double or treble a farmer's yield, even if the overall season is drier than average.

Forecast information needs to reach users in a more timely manner. Several users noted that they receive ENSO forecasts two months after the fact. By that time, they may have lost some opportunities to act. The impacts of the event may have already become visible, making the impacts information useful only to assess the validity of the forecast. Many users in the region also noted that they do not receive information on a regular basis. This makes it difficult to monitor the progress of the ENSO cycle.

Some users felt that the information provided on ENSO (e.g., forecasts, remote sensing data) is not readily accessible. In many cases, users are unfamiliar with the intricacies of the ENSO phenomenon and may be confused by information on SSTs, the SOI and OLR (out-going longwave radiation). And several users have noted that unanalyzed raw data is of little use to them. It would be helpful if such data were to be accompanied by interpretation in order to help the user to better understand the implications of what has been observed.

7.5. Conclusion

Our research strongly suggests that there is a need for educating potential and even actual users on the availability, confidence limits, and ways in which ENSO-related information, (forecasts as well as "ENSO climatology") can be applied to decision making. Good scientific information must be presented to potential users within a timeframe that enhances the possibility that it will be used in decision-making processes and with enough reliability that decision makers will feel confident when they take it into consideration. It is also clear from the assessment that a formal mechanism must be developed in the region for the analysis and dissemination of ENSO-related information. While ENSO information may only be one piece of information out of many that a decision maker must evaluate, it should be on the list of information to be evaluated whether or not it proves to have been important in that particular instance.

  |   Previous Chapter   |

  |   Main Page   |   Table of Contents   |   Acronyms   |   References   |