Geospatial and Statistical Modeling of Artisanal Mining Populations in Kenieba, Mali

Abstract

Artisanal and small-scale mining (ASM) has been both a formal and an informal part of many developing economies for decades. It is estimated that 13 to 20 million people are involved in artisanal mining globally. While it has been a prevalent part of poverty alleviation strategies worldwide, the people involved in ASM are a largely understudied population. This study uses a combination of remotely sensed data and fieldwork questionnaires, compiled and analyzed in a GIS, to model the spatial demographics of the ASM population in Kenieba, Mali. This paper and map discuss and illustrate the regional population of Kenieba, the number of miners per mine and the spatial-temporal movement patterns and proximity to mine sites.

The Kenieba study area is located in western Mali along the Senegalese and Guinean borders. At most sites both gold and diamonds are being mined, as the mining of gold is a more profitable activity in this region than diamond mining. The map features three, 3-D figures representing land cover topography and population with the results of a path-distance analysis that display the time it would take an artisanal miner on foot to traverse the terrain of the study area. Due to the 3-D, oblique orientation of the map the scale varies throughout the figures.

Keywords:

• geospatial
• statistical
• modeling
• West Africa
• Mali

1. Introduction

Artisanal and small-scale mining (ASM) involves the use of simple tools to extract near-surface mineral ore and often occurs informally or illegally throughout many parts of the developing world. ASM provides participants with an opportunity for rural development and poverty alleviation in developing regions where few alternative exists; however, little is known about artisanal mining communities. Conducting research on miners is challenging because mining activities are seasonal, they are often conducted outside the legal and regulatory framework, workers are typically transitory, and the dynamic market conditions for mineral commodities, such as gold and diamonds, cause people to enter and leave the sector rapidly. Some estimates of the global artisanal mining population are as high as 20 million (Vlassenroot & Van Bockstael, 2008). However, few studies have collected data on the spatial trends of ASM communities, leaving a significant gap of baseline demographics and spatial awareness of their livelihoods (Hilson & Maponga, 2004).

ASM has many social and environmental effects that perpetuate the marginalization of artisanal miners. Working conditions are poor, the work is physically demanding, and there are often no regulations to ensure safety at the mine sites. Young children are also often engaged in part of the mineral recovery process, violating many international child welfare standards (International Labour Organization, 2007). Disease and sexual abuse are rife at some mine sites as well. However, research shows that the sector provides much needed seasonal work (miners often farm during planting and harvesting seasons) and when women are involved in mining and wage earning their standing in household affairs can improve (Bøas & Hatløy, 2008). Policy makers are divided as to whether ASM should be regarded as a ‘sustainable livelihood’ or whether ‘alternative livelihood’ ventures should be implemented instead. If considered a ‘sustainable livelihood’ ASM should be supported and formalized, whereas if an ‘alternate livelihood’ is pursued then advocates suggest transferring workers from mining to other livelihoods, such as fish farming (Jønsson & Fold, 2011).

Land degradation and mercury contamination are the largest environmental problems associated with ASM. There are no standards for starting and shutting down mine sites, and deforestation, erosion, and contamination have become rampant problems in mining regions. Abandoned pits that are not filled-in accumulate water and become breeding grounds for mosquitoes, increasing malaria transmission. Miners also may use mercury to assist with the recovery of gold through the process of amalgamation. Mercury is a highly toxic chemical, and miners often handle it directly, without using gloves or other protective measures. As a result, it can be absorbed into the body throught the skin and affect the nervous system (Hilson, 2002). One study conducted in the Birrimian areas of Mali (including the Faleme River basin, which is within the study area examined in this paper) showed that copper, manganese, iron, Total Dissolved Solids, and sulfate ions were at concentrations above the World Health Organization's limits, especially in areas near gold mining activities. The areas with the highest contamination levels were found in shallow wells (Bokar et al., 2011).

Artisanal mining has occurred in the Kenieba region of Mali for over 2000 years. This area contains primary and secondary (alluvial) sources of both gold and diamonds. Gold has been mined by artisanal miners from veins, primarily in the Tabakoto area, as well as from ancient and recent alluvial deposits (Chirico et al., 2010). These deposits have been exploited from the surface to a depth of up to 50 meters using ASM methods. Diamonds were discovered in Mali in 1954, and in 2003 Mali was the third largest exporter of gold in Africa (Bokar et al., 2011). To date, a variable number of kimberlites have been reported in Mali, eight of which are commonly referenced as having been sampled positive for diamonds. However, none of Mali's kimberlites are currently being mined. Rather, alluvial diamonds are recovered as a secondary product of alluvial gold mining (Chirico et al., 2010).

No census has been conducted of the ASM population, resulting in a lack of concrete data and the marginalization of this sector. Fieldwork and visualization in a GIS can provide detailed information and analysis which can be used by policy makers, NGOs, and civil society organizations. The maps and visual representations presented in this study provide an estimate of the number of people involved in mining in Kenieba, Mali, the number of people living in the surrounding areas, and the routes traveled by workers to reach mine sites.

2. Data

This map uses questionnaire1 information gathered during geologic fieldwork conducted by USGS personnel in February and March of 2007. The questionnaires contain information about the mining sites (geology and lithology of the mine sites and data on the extracted material) and certain demographic data on the mineworkers themselves. A total of 33 mine sites were visited in the Kenieba study area during fieldwork. Information collected during fieldwork, such as where miners live and how they travel to and from mine sites were essential to the analysis conducted in this study (USGS. Chirico, 2009).

Oak Ridge National Laboratory's (2006) Landscan dataset, a remotely sensed dataset showing world ambient population at night, was used to calculate the population of the region. To model the terrain, a 90 meter resolution digital elevation model (DEM) derived from the Shuttle Radar Topography Mission (SRTM) (NASA, 2009) was used. The European Space Agency's (ESA) (2005) GlobCover raster dataset was used to map land cover, and shows 21 types of land cover classifications. For the purposes of this analysis, the raster dataset was reclassified to show the seven most common land cover classes within the study area.

3. Methodology

The methodology was comprised of the following steps: GIS database development, network analysis, population analysis, and map creation.

3.1 GIS creation

Data layers representing the locations of mines, villages, rivers and roads data were developed from fieldwork data, Joint Operation Graphics (U.S. National Imagery and Mapping Agency) topographic map sheets at 1:250,000 scale, and satellite imagery, and were clipped to the study area extent.

3.2 Path distance analysis

Once the basic GIS for the study region was created, the path-distance function was used to analyze the time it would take a miner to travel on foot within the region from a mine site to the surrounding area. Path-distance calculates the accumulative cost over a cost surface, compensating for the actual surface distance that must be traveled and for the horizontal and vertical factors, which influence the total cost of moving from one location to another. In this analysis the ‘cost’ is the time, in minutes, needed to travel, on foot, from a mining area across the 1000 × 1000 meter cell. Roads, rivers, and land cover data were modeled in the terrain cost data layer, which is then used in the path-distance analysis. Based on walking speed experiments conducted by the authors, it was determined that the average person can walk 5.5 kilometers per hour (km/h) on a paved surface. For all road types, rivers and land cover classes, estimates of speed were determined and values were derived from between 5.5 and .05 km/h based on terrain type (equation: 60 minutes/Speed = Value).

The path-distance model was created to show the amount of time needed to traverse across a 1000 square meter cell on foot, depending on the conditions on the ground, such as the location of roads, rivers, relief, and type of land cover. The results were then symbolized to reflect one hour interval ranges (from up to one hour to eight hours and above).

3.3 Population analysis

Due to the lack of artisanal miner census it is difficult to estimate the number of people involved in ASM, and in turn the number of people living near mine sites with livelihoods that are dependent on the mining sector. Using the remotely sensed population data (Landscan), an estimate can be made of the number of people living in the region to determine if the areas of high population are located in proximity to mine sites.

Using the path-distance zones and the Landscan data, an understanding of ambient population as a whole and for the population within each hour range can be determined.

4. Results

Table 1 displays the union of the path-distance analysis and the population analysis. For each path-distance hour range, the population and percentage of total population is given. This allows for an understanding of the number of people living in proximity to mining areas and the distance which people in each zone must walk to reach a mine area. The largest population is found within the ‘8 hours and Up range’ because this is the largest geographic range. Similarly, the smallest population is found within the smallest geographic range, ‘Up to 1 Hour’. However, the second and third largest populations are within the ‘1–2 Hour’ and ‘2–3 Hour’ ranges, respectively, and the second smallest population is within the ‘7–8 Hour’ range. Therefore, the size of the ambient population is not necessarily correlated to the geographic size of the ranges; rather, it is correlated to their proximity to mining areas. This leads to the conclusion that people in the region either choose to settle relatively close to mine sites or that mine sites develop in their specific locations because of their proximity to populated settlements.

Table 1. Ambient population within study area.

Hour Ranges	Ambient Population in Hour Ranges	Percent of Total Population (%)
Up to 1 Hour	6002	3.33
1–2 Hours	20,619	11.46
2–3 Hours	14,656	8.14
3–4 Hours	11,840	6.58
4–5 Hours	9691	5.38
5–6 Hours	9126	5.07
6–7 Hours	11,008	6.58
7–8 Hours	8839	4.91
8 Hours and Up	88,218	49.01
Total Population	179,999	100

Using fieldwork questionnaires collected at 33 mine sites, it was determined that there are approximately 9,000 miners in the Kenieba study area. That figure is 21.7% of the population living within 3 hours walking distance (41,277 people) from mining areas, 12.5% of the population living within 6 hours (71,934), and 4.98% of the total population for the study area (179,999) (Calculated from Table 1). 21.7% is a significant portion of the immediate population to be working in ASM. Jønsson and Fold (2011) remark that miners in a spatially concentrated area can bring opportunities for small businesses catering to miners’ needs. Such businesses include stores selling mining tools or items for personal use, food stands with produce or prepared meals, as well as bars and brothels. This leads to the conclusion that many more than 9,000 people are indirectly involved in the mining sector in this region. Including those in ancillary tasks and industries would increase the estimated number of people actively involved in the ASM sector in this region.

The total population of the region is calculated to be approximately 180,000 (Table 1). Landscan is a highly regarded dataset, but the data can experience a large range of percentage differences between the ambient population detected by Landscan and available census data in tested locations. According to Dobson et al.'s (2000) examination of the Landscan dataset, this difference can range from −14% to +8%. The 2009 regional population of Kenieba, calculated from the National Institute of Statistics of Mali, is estimated to be 194,153 (INSTAT). The percent difference between the two figures is 7.56%, which falls within the range listed by Dobson et al. (2000). One possible explanation for this difference in population counts is that the population could simply have increased by 14,154 (the difference between the two figures) from 2006 to 2009.

5. Conclusion

This map highlights the relationships between the locations of populations, mine sites and the mine workers, as well as the pedestrian travel times between mine sites and surrounding villages based on the topography and land cover characteristics of the study area.

This methodology can be adapted to estimate population and travel times in other ASM regions. The results of this study would be useful for civil society organizations in Mali focusing on issues related to mining, land degradation, mercury poisoning, or general health services. The information generated from such analysis could be used, for example, to determine the locations in which services targeting artisanal miner populations should be placed to ensure maximum impact on this population.

Software

ESRI's ArcMap v.102 was used to carry out the population and travel time distance analysis and Golden Software's Surfer (v10.0)² was used for 3D modeling. The results of all analysis were then imported into Adobe Illustrator CS 5² for further cartographic layout and design.

Notes

The fieldwork conducted in Mali was undertaken under the auspices of the U.S. Department of State in support of the Kimberly Process. The goal of this fieldwork was to determine the diamond resource endowment and production capacity of Mali. Geologic and geomorphic data were collected to determine the extent of the country's diamond deposits. This data was then used in combination with socio-demographic data, such as the number of miners working at individual sites and the tools and methods employed, to determine the number of diamonds Mali is capable of producing per year. A comprehensive questionnaire was developed and used by the research team to gather this data. Geologic information was acquired through observation and site measurements, while socio-demographic data was largely collected by conducting individual and group interviews with miners.

ESRI's Arc Map, Golden Software's Surfer and Adobe Illustrator were all used in the creation of this map. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.