The Geography of Clean Water for All
How high resolution population data and GIS are helping reach the most vulnerable communities
Our Roadmap to Impact
Our Roadmap to Impact: World Vision’s 2021-2025 Global WASH Business Plan sets forth World Vision’s five-year strategy and organizational commitments in the water, sanitation, and hygiene (WASH) sector. More specifically, the Global WASH Business Plan describes how World Vision will advance our mission to bring transformative WASH services to every child we serve across more than 40 countries. The Business Plan articulates four overarching goals, six core programming areas, and four program quality initiatives. In this regard, Goal 1 of the Global WASH Business Plan is:
Accelerate universal and equitable access to water, sanitation, and hygiene services in support of Sustainable Development Goal 6
Against this backdrop, World Vision is seeking to strengthen our WASH monitoring approaches and standards, adopting best-in-class tools and practices that bolster program quality, generate evidence of impact, engender sustainability, and deliver on our promises to communities and donors alike. Toward this end, there are many evolving technologies that can help us to better monitor progress and plan new investments, including: digital survey applications for field data collection; high resolution satellite imagery for remote data collection and validation; and geographic information systems (GIS) for rigorous data analysis and visualization. While there are a host of examples to consider, this story map explores an especially promising tool for accelerating universal and equitable access to WASH.
This storymap is the first of World Vision's series on the "Geography of Clean Water for All". The series explores the role of geospatial data, digital tools, and catalytic partnerships as World Vision strives to reach everyone, everywhere we work with clean water. Additional series stories are: A Promise Verified .
High Resolution Population Data: Putting Vulnerable Communities on the Map
The ambitious target of SDG 6 of " water and sanitation for all " requires a deep understanding of the geography of the phrase "for all". In other words, reaching "all" mandates a deep dive into the distribution of the population across the landscape, including the most remote communities that are often not visible on the map. While it may come as a surprise, the fact is that knowing where people live and how many people live there continues to be a significant challenge in humanitarian and development assistance, including in the WASH sector.
Enter the powerful combination of satellite imagery, machine learning, and GIS! To help fill this critical data gap, teams of earth scientists, remote-sensing specialists, data scientists, geographers, and everyday map enthusiasts are coming together to create high resolution population analyses, helping us to map not just cities and towns but rural villages and even individual households. Fueled by the increased availability of detailed satellite photos, advances in computer learning, new forms of online collaboration and crowd-sourcing, exciting datasets have emerged with unprecedented detail on where people live. While the options are continually expanding and evolving, three datasets that we often turn to are:
Figure 1: Building footprints are merged with population statistics ( Source )
Facebook Data for Good Population Maps: Facebook's Data for Good (D4G) team has partnered with Columbia University's Center for International Earth Science Information Network (CIESIN) to create detailed population estimates for 30 by 30 meter cells across a global grid. The D4G team uses machine learning to detect buildings from satellite imagery while engaging CIESIN to overlay population statistics across the mosaic of settlements. Per Figure 1, the presence (or absence) of a building footprint combined with census data results in an estimate of the population for each cell across the grid. The complete data set is available by country on the UN Humanitarian Data Exchange .
Figure 2: The population distribution of Kumasi, Ghana using WorldPop's Open Population Repository ( Source )
WorldPop Open Spatial Demographic Data and Research : Led by demographic experts from Southampton University, the WorldPop program has developed open, high-resolution geo-spatial datasets for population distribution, spatial demographics, and key development indicators. WorldPop maps incorporate a variety of remote sensing data types---including settlement locations, land cover, roads, nightlight luminosity, and topography---to create 100 by 100 meter gridded population models (see Stevens et al, 2015 for greater detail on the methodology).
OpenStreetMap (OSM) : Finally, OSM has empowered users from every corner of the globe to contribute location-based data in amazing ways. The Humanitarian OSM Team (HOT) and related initiatives such as Missing Maps have mobilized an army of virtual volunteers to digitally trace buildings, roads, and other critical infrastructure using high resolution satellite imagery. The result is extraordinary detail of the built environment. Figure 3 below, for example, shows the city center of Kumasi, Ghana where about 2,200 buildings were digitized in OSM as of 2018 (per image left of slider). The image to the right of the slider, however, shows a sample of the more than 83,000 buildings added to the map of Kumasi by 2021. Move the slider back and forth to compare the significant progress achieved by OSM mappers.
Figure 3: The city center of Kumasi, Ghana in OpenStreetMap in 2018 (left) and 2021 (right). ( Source )
The View from Above: The Role of Satellite Imagery
Figure 4: Satellite imagery of a community in Nyamagabe, Rwanda
While the three datasets use different models and digital tools, they also have a common critical ingredient: high resolution satellite imagery. Simply put, dramatic increases in the availability and accessibility of earth observation imagery have transformed population mapping.
As an example, Figure 4 provides a typical, high-level view of settlements in Nyamagabe District in southern Rwanda. In comparison, in the center of Figure 5 we see the same view but with darkened, pixelated areas where pattern recognition programs have detected built structures, providing immediate insights into population density across the landscape. In other words, the darkened, pixelated areas are based on hundreds—if not thousands—of satellite images where the telltale patterns of houses, schools, and markets were detected, as shown in the group of images at left. Meanwhile, the unselected, undarkened areas indicate that only natural surfaces and objects were visible in the satellite imagery, such as those shown in collection at right.
Figure 5: Comparison of satellite imagery with (at left) and without (at right) building detection ( Source )
In this way, the remote sensing images captured by orbiting satellites and computer-aided pattern recognition programs can take us from space to village, helping map where people live even in the most distant communities. But how does this help us in the WASH sector? Let's consider two examples.
Tracking Progress in Rwanda
Active in Rwanda since 1994, World Vision began scaling up its WASH program in support of the Government of Rwanda in 2012. Scroll through the following maps to learn more about how population data is helping World Vision track progress towards universal service coverage in areas where we work.
World Vision's WASH programming spans all of Rwanda's five provinces, including at least 40 sub-districts (shown in the map at right) across 13 districts. World Vision works closely with each local government where we have programming to design, plan, and co-finance new water systems. For this illustration, we will focus on the District of Nyamagabe in the Southern Province.
Figure 6: A new tap in Southern Province, Rwanda
Before looking at two specific sub-districts where World Vision has been working over the past five years, let's first consider the distribution of the population across the entirety of Nyamagabe District. Per the map at right, the district is composed of 17 sub-districts.
We see here the “old way" of visualizing population, with each sub-district classified by its aggregate population using a color ramp. The Nkomane Sub-District in the northwest corner, for example, has an estimated population of about 20,000. This view, however, misses important details...
In contrast, here is the New Way using the high-resolution population density data. What becomes suddenly obvious—which was hidden before—is that the 20,000 people living in Nkomane Sub-District all live in the eastern half of this administrative area. Indeed, this view shows that very few people live along the western edge of Nyamagabe as a result of a national park. Now, let's take a closer look at two sub-districts in the southeast corner: Gasaka and Kamegeri.
World Vision previously partnered with Gasaka and Kamegeri to inventory water supply facilities across both sub-districts, providing an ideal data set to illustrate the use of high resolution population data. Notably, the combined population of Gasaka---home to the district capital---and Kamegeri is about 63,000 people.
Now, let's add several layers of data to better understand the geography of water access in these locales...
First, let's bring in our high-resolution population density layer (which, for this illustration, is the Facebook/CIESIN population data discussed previously). Immediately we get a better sense for where the 63,000 residents of these two sub-districts actually live, including areas with more concentrated populations (e.g. northeast Gasaka) as well as the less populated areas (e.g. south Kamegeri).
Now, let’s overlay our historical data on water collection point locations...
With this layer, we can see a series of water collection points across the two sub-districts, including those supported by World Vision, local government, the private sector, and other implementing partners. By comparing the locations of these water points to the population layer, we begin to get a better idea of water service coverage levels. But let's take it one step further...
…By adding a 500-meter radius around each water collection point. Importantly, the Government of Rwanda’s policy is that improved water collection points should be no more than 500 meters from each rural household. With this policy in mind, as a final step, we can tap into the power of geo-spatial analysis and run a calculation of the number of people living within 500 meters of all water points depicted.
This analysis tells us that, at the time the data was collected, about 60% of the population lived within the government’s prescribed distance of a water collection point. Importantly, this type of illustration can also help bring attention to areas that need further investigation, such as the potentially under-served or unserved areas encircled by dashed ovals. It would be important, in other words, to talk with communities and local government about such areas which may not yet have ready access to clean water.
Planning for Universal Service Coverage in Zambia
First launched in 2008, World Vision's WASH Program in Zambia helps to improve health, nutrition, and education outcomes through universal access to sustainable and safely managed WASH services. World Vision dramatically scaled up our Zambia WASH programs in 2011 and, since that time, has helped the Government of Zambia extend clean water services to approximately 1.5 million people. Building on this work, World Vision Zambia is now planning to support universal service coverage in priority areas over the coming five years. Our second illustration looks at how population data is informing the planning of this critical effort.
In early 2021, World Vision Zambia conducted an assessment of access to clean water across more than 100 wards shown here. Our teams visited 10,000+ water points, 630 schools, and 195 health facilities, using digital survey tools to record and geo-reference data. Importantly, after mapping the resulting field data and overlaying population density, the initial analysis showed that nearly half of the population in surveyed wards lived beyond 500 meters from a safe, functional water source.
The geospatial analysis of the field survey and population data also showed significant disparities amongst the wards, including 21 wards where more than 75% of people live beyond 500 meters from clean water facility. One such ward is Kasenga, located in the Mwamba Area Program in Northern Zambia. Shown at right, Kasenga Ward had one of the highest proportions of people living beyond 500 meters of any area surveyed. Let's take a closer look at the geography of water services within this jurisdiction...
The Geography of Water Services in Kasenga, Northern Zambia
We begin with the boundary of the ward as well as the locations of schools and healthcare facilities. The total population of the ward of about 26,000 people, and let’s bring in this population density layer now...
Utilizing again the Facebook/Columbia University data, we can immediately see that the population of Kansenga---represented by the clusters of black pixels---is highly distributed around the ward. Even the small villages of Ngosito, Chilongoshi, Misengo, and Kakosa show low population densities. Now let's add in the water point data collected by our Zambia team using the digital survey tool mWater .
The World Vision team visited communities across the ward to observe their primary water source, surveying more than 250 water collection sources. Of those visited, almost 60% were unimproved (or unsafe) while just over 40% were improved water points. Amongst the improved water points, however, about 19% were also not fully functional at the time of the visit. The following chart summarizes the results.
Similar to the example in Rwanda, we can assess the proximity of the population to an improved water facility using GIS. What we find is that 77% of the population—or nearly 20,000 people—live beyond the prescribed distance to an improved water point, making this ward one of the greatest areas of need when it comes to clean water access. Also, as seen in the cluster of improved water points in southeastern Kasenga, the majority of people with easy access to an improved source live in the village of Ngosito.
Excluding areas already covered by improved, functional water points, we can use a heat map analysis to begin to identify specific communities for potential investment. We can see, for example, unserved areas near the villages of Ngosito, Chilongoshi, and Kakosa. Misengo Village also only has a single functional water point. Further, we see multiple clusters of households distant from the main road that are completely unserved.
Indeed, a spatial pattern visible in Kasenga is the tendency of improved water points to be located along the main road. In this regard, the World Bank’s rural access index—which serves as an indicator for Sustainable Development Goal 9—uses distance from an all-weather road to help identify vulnerable communities. The benchmark for the RAI is 2 km. In Kasenga, there is just one functional, improved water point beyond 2 km from an all-surface road (located at Misengo health clinic). This is despite the fact that 60% of the population of the ward live beyond 2 km.
In closing our visit to Kasenga, let's take a brief tour of several of the water collection sources photographed by the World Vision Zambia team during the survey.
The Geography of Water for All: Making the Invisible Visible
Mapping the Unmapped
Facilitating universal service coverage requires a comprehensive understanding of the geography of the WASH landscape, or the way in which water supply services are distributed across a district. High resolution population datasets represent an important tool in understanding that geography, especially in rural areas. In a sense, these datasets help us to map the unmapped by revealing the locations of even the most remote villages, such as the community where Tiness and her baby reside in southern Zambia (photo at left).
Thus, as we partner with governments and communities alike, World Vision will continue to plan investments and monitor progress using these important tools and technologies, making the invisible visible to ensure that no one is left behind in the effort to bring clean water to all.
To learn more about World Vision's work and partner with us, just follow the link below.
