This post introduces Philip Gooding and Nadia Fekih’s recently published Journal of Southern African Studies article, “A Climate History of Early Dutch Settlement at Cape Town, 1652–62.”

---

An ongoing trend in environmental history is to use climate data from archival sources to recontextualize historical events against the backdrop of climate variability. Such is the aim of the Appraising Risk project, an interdisciplinary SSHRC Partnership hosted at the Indian Ocean World Centre, McGill University, for which I am a research assistant. Building on this body of scholarship, my research uses mid-seventeenth-century Cape Town as a case study. The advent of Dutch colonization from 1652 coincides with the height of the Little Ice Age, a global cooling event that underpinned what Geoffrey Parker refers to as a “global crisis” of intense political instability.¹ Using data collected from the Journal of Jan Van Riebeeck (hereafter: the Journal), I used interdisciplinary methods to digitally visualize daily weather reports at the Cape, and found evidence of deviations from typical seasonal wind and rainfall trends. The consequences of these anomalies have been discussed in a recently published collaborative article in the Journal of Southern African Studies. In this blog post, I visualise some of the meteorological data I gathered and standardised to create this article.

The Journal is a three-volume compilation of documents in the Dutch East India Company (VOC) Archives from the first decade of Dutch settlement at the Cape of Good Hope, 1652-1662.² Sometimes written by the station’s commander Jan van Riebeeck (1619-1677), it provides an account of the Dutch VOC’s daily operations at the Cape and its relations with Indigenous populations. Each diary entry includes a description of daily weather conditions, including reports on wind and rainfall. A typical series of reports is as follows:

1656

January 6th.—Fine weather; light breeze from N.W

January 7th.—Ditto.

January 8th.—Ditto. Wet weather in the morning.

January 9th (Sunday).—Dry. S.S.E. breeze.

January 10th.—Weather as yesterday.

On occasion, the Journal also suggests ways in which climactic conditions affected the settlement:

January 13th.—S.S. East wind and cloudy sky.

January 14th.—Westerly breeze, and clear sky.

January 15th.—Dark sky, wind N.W.     16th, the same.

January 17th.—As above, dark sky and a dry west wind, destroying the pasturage so that nothing almost is left for the cattle to eat; some old cows are failing in consequence, and we are compelled to kill them for food; the milk [sic] cows are drying up; every year it becomes plainer that during the dry season very little can be obtained from them.

The Journal, therefore, provides primary data that suggest how the first Dutch settlers contended with the enviro-climatic conditions of the early Cape Colony.

In total, the Journal contains over 4,000 qualitative weather reports over a 10-year span. The first challenge for our collaborative research was to standardise these reports to assess climate variability and change over time. For wind observations, I borrowed the Journal’s categorization system. The reports always denote wind by the direction from which it comes. Its strength, if specified, is described in one of three ways: “Strong,” “Slight,” or “Variable.” Entries without strength specification are classified as “No Data.” Rainfall, however, was not catalogued as consistently. Based on its description, I sorted each report of rainfall into one of four categories: “Heavy,” “Light,” “Intermittent,” and “No Description.” Additionally, I tabulated the number of rainy days in each month and calculated it as a percentage of the present-day monthly average of rainy days in Cape Town. This provided a way of assessing the amount of rainfall in each month, compared to modern averages. The standardized data from this project is available open-access in the McGill University Dataverse.

The standardized data were then used to create two types of visualization: 1) graphical representations; and 2) an animated map using QGIS’ time series function. The graphical representations display trends of wind direction and rainy days, focusing on the wet season (June-August), which is where we found most anomalies. Figure 1 displays the number of wind reports per season by direction of provenance; figure 2 shows the number of recorded rainy days in the wet season for each year compared to the seasonal average.

Figure 1: Wind Reports by Direction at the Cape of Good Hope during the Wet Season (June-August), 1652-61; Figure 2: Number of Rainy Days at the Cape of Good Hope per Wet Season (June-August), 1652-61.

In the time-series map, daily wind, rainfall, and drought data are represented by separate shapefiles whose properties vary according to their assigned attributes. For instance, wind is represented by yellow arrows oriented in the direction from which it comes, while their size depends on their relative strength. “Strong” winds are denoted by a thirty-point arrow, “Variable winds” by a twenty-point arrow, and “Slight” winds by a ten-point arrow.  Likewise, rainfall is denoted by blue diamonds which vary in size on the same scale. Grey symbols for wind or rainfall indicate that there is no description of their strength provided in the journals. The degree of drought (relative to monthly modern averages) is represented by a large square, which changes between shades of red and blue depending on the degree of drought or excess of that month. Red indicates rainfall deficit, while blue indicates a surplus.  

These visualizations suggest that, in the mid-seventeenth century, the climate deviated from patterns typical of Cape Town. Southern Africa’s wet season (June-August) is normally characterized by predominantly northwesterly winds that bring about moist air and rainy conditions. However, as Figure 1 shows, NW winds only made up around half of the wind reports in any given year between 1652-61. This suggests that the winds during this decade were uncharacteristically variable, thus bringing less rain to the region. This fact may explain why, as shown in Figure 2, the Journal suggests that there several seasons of below-average rainfall in from 1656 onwards, except in 1660.

Figure 3 provides a way of accounting for the role of wind strength in these climate anomalies. The dynamic nature of the time-series allows wind arrows to take on multiple attributes, i.e., direction and size. Additionally, because the wind data is directly overlain on the drought/excess data (the colour-shifting square), the time-series reinforces the correlation between weaker/variable winds and drought. For the wet seasons during the period under review, the orientation (direction) of the wind arrow is highly variable from day-to-day, while its size (strength) is rarely described as “Strong.” In the latter half of the decade, these wind conditions are overlain on the square in varying shades of red, indicating drought.

The goal of digitizing and visualizing these data is to contribute to the growing body of work being done to reconstruct past climates and analyze how climactic change can impact societies, past and present. Graphical and GIS representations facilitate this work, as they provide an accessible way to interpret the patterns found within thousands of pages of primary data. As we discuss in more depth in our collaborative article, climatic conditions during the first decade of Dutch settlement at what is now Cape Town–abnormal as they appear to have been–significantly affected many of the key historical threads of this period, including those related to food production, commerce, disease, and political stability. We urge interested readers to consult the full article, and to contact the authors if they require access.

Nadia Fekih is a Research Assistant at the Indian Ocean World Centre, McGill University.    

---

Notes

¹ Geoffrey Parker, Global Crisis: War, Climate Change and Catastrophe in the Seventeenth Century (New Haven: Yale University Press, 2017).

² Jan van Riebeeck, Journal of Jan Van Riebeeck, ed. Hendrik Bernardus Thom (Cape Town: Balkema, 1952).