ENSO Africa

In the absence of oral histories, analyses of historical documents held in missionary and colonial archives offer an invaluable opportunity to fill knowledge-gaps about societal responses to extreme climate events. Documents describing, for example, social change, technological adaptation, narrative and myth with respect to rainfall variability can be used to investigate how different sectors of society responded to, and articulated knowledge about, periods of unusual climate. When interpreted within the wider context of social, economic, political and demographic changes over time, such information can provide a guide to where the most critical sensitivities to future climate change may lie.

In addition to their potential for providing insights into climate impacts upon society, documentary sources also contain valuable direct and indirect information on weather that can be used for climate - especially precipitation - reconstruction. The majority of climate models for southern Africa depend upon instrumental meteorological data. However, with the exception of parts of South Africa, these data have only been collected systematically since the late19th century. Document-derived weather information can be used to extend the instrumental record and, critically, identify climate extremes. The resulting chronologies of climate variability can be used to form a meteorological framework within which any societal changes can be viewed.

The majority of documentary sources relating to Madagascar, Malawi and South Africa that contain references to climate stem from European missionary or colonial archives. These include materials which collectively span the 19th century, written by missionaries, explorers, botanists and government officials.

Among the most important are the collections of unpublished documents preserved within archives in Aberdeen, Birmingham, Edinburgh, London, Oxford (UK), Coromandel (Mauritius), Berlin, Hermannsburg (Germany), Boston (USA), Durban and Pietermaritzburg (South Africa), Lilongwe (Malawi) and Stavanger (Norway).

Download the list of documentary sources (pdf).

The project ran from March 2010 to August 2013.

The overall aim of the ENSO Africa project was to examine the socio-economic consequences of, and human responses to, climatic extremes associated with historical El Niño events during the 19th century in three case study areas: Madagascar, Malawi and KwaZulu-Natal.

The three areas have been selected because they all: 

• lie within the area of greatest ENSO-sensitivity in southern Africa
• experienced significant and well-documented ENSO-related modulation of summer rainfall during the 20th century (and by inference during the pre-instrumental period)
• have limited or no available climate information prior to the instrumental period
• had contrasting colonial and pre-colonial socio-political structures
• benefit from a considerable volume of 19th century documentary materials through which to reconstruct climate histories and identify any social repercussions.

We planned to meet our overall aim by addressing the following objectives:

1. To identify the spatial and temporal evolution of climate extremes associated with known 19th century ENSO events in the three study areas using historical documentary sources combined with climate modelling;
2. To identify the scale of impact and nature of human responses to these extreme climatic conditions using historical documentary sources;
3. To examine how societal responses to climate variability may have differed during the 19th century, and explore any socio-political changes that may have influenced resilience and vulnerability during this time period.

All 19th century materials in each of the archive collections were scrutinised for accounts of climate-related conditions and any associated societal repercussions. The types of climate-related information identified included descriptions of droughts, storms, floods and other extreme weather events, accounts of the onset, intensity and duration of seasonal rains, plus less direct indicators of rainfall variability such as relative harvest yields. Instrumental weather observations not captured previously by the meteorological services of Madagascar, Malawi and South Africa were also noted. All descriptions of related societal responses to harvest failure, food scarcity, economic dislocation, disease and social strife that may have been climatically driven were recorded. These included local and regional population dispersals, settlement relocation, technological innovations and coping strategies, plus commentary on climatic narratives, customs and environmental myths.

To date, the project has generated significant new data for understanding the patterns of, and controls upon, past climate variability. Outputs have focussed on historical tropical cyclone activity (Nash et al., 2015) and rainfall variability (Hannaford and Nash, 2016; Nash et al., 2016) in Madagascar and KwaZulu-Natal, respectively. We are in the process of constructing an equivalent assessment of historical rainfall variability for Malawi, and are exploring the societal implications of climate variability in all three of our case study areas.

Historical documents from Madagascar have been used to identify the damage caused by tropical cyclones during the late 19th century. Maps of damage have been used to reconstruct the pathways of individual tropical cyclones across the island (see Figure 2), and accounts of the extent of damage used to estimate cyclone intensity. We were able to reconstruct the occurrence and approximate pathways of 20 major tropical cyclones between 1862 and 1900, of which only 17 have been recognised previously. Our results suggest that fewer tropical cyclones made landfall on Madagascar during the 19th century compared with the period since 1970. Our results are now being fed into international databases such as IBTrACS.

Historical sources from former Natal and Zululand proved to be incredibly rich in detail, and have been used to develop the first combined seasonal and annual rainfall reconstruction for anywhere in Africa (see Figure 3). The reconstruction identifies eight severe or multi-year droughts (the rainy seasons of 1836–38, 1861–63, 1865–66, 1868–70, 1876–79, 1883–85, 1886–90 and 1895–1900), the most severe of which was the drought of 1861-63 and the most prolonged that of 1895-1900. A further six severe or multi-year wetter periods were also identified (1847–49, 1854–57, 1863–65, 1879–81, 1890–91 and 1892–94). The early 1890s included a run of extremely wet years and appears to have been the wettest period of the 19th century, while the 1850s was the wettest decade overall.

The results of the research have been combined with other annually-resolved reconstructions of moisture variability from the region, and used to produce a multi-proxy record of rainfall variability in the southern African summer rainfall zone for the last 200 years. This reconstruction shows that the 19th century was, on average, wetter than the 20th century. It also identifies a long-term decline in the level of precipitation in the summer rainfall zone that has continued into the 21st century. El Niño events plus variations in sea surface temperatures in the Indian Ocean are identified as the biggest contributors to regional rainfall fluctuations over the last 200 years.