Mapping Kenya’s Coast: Abigail’s Journey with Digital Earth Africa

Meet Abigail Kagema, a Geomatic Engineering and Geospatial Information Systems graduate from Jomo Kenyatta University of Agriculture and Technology (JKUAT), whose research bridges GIS, remote sensing, and Earth Observation analytics. Passionate about applying machine learning to real-world environmental challenges, Abigail focused her undergraduate thesis on coastal vulnerability along the Kenyan coastline, developing a data-driven Coastal Vulnerability Index that combines physical and socioeconomic indicators. By leveraging Digital Earth Africa’s tide-normalised satellite datasets and analysis-ready tools, she generated a consistent 24-year shoreline history, uncovering critical insights into erosion hotspots and informing adaptation strategies. Abigail’s work highlights how accessible, high-quality EO data can empower researchers to address local environmental risks with precision and impact.

Briefly describe your academic background and research focus.

I hold a Bachelor of Science degree in Geomatic Engineering and Geospatial Information Systems from Jomo Kenyatta University of Agriculture and Technology (JKUAT). My academic interests sit at the intersection of GIS, remote sensing, and Earth Observation analytics, with a particular focus on machine learning, natural resources, climate resilience and coastal risk.

For my undergraduate research I specialised in coastal vulnerability modelling along the Kenyan coastline. I combined physical indicators (such as shoreline change, elevation, slope, waves, and sea-level trends) with socioeconomic data to derive a Coastal Vulnerability Index that can guide adaptation and planning.

How did you discover Digital Earth Africa, and what drew you to its tools?

I first heard about Digital Earth Africa during remote sensing class projects, where we were introduced to different satellite data platforms for the African continent. However, I only started working with DE Africa directly when I began my Bachelor’s thesis.

While looking for a way to obtain satellite-derived shorelines of the Kenyan coast at approximately 0 m above mean sea level, I came across the masking and tidal normalisation approach used by Digital Earth Australia. Further reading led me to the DE Africa Coastlines product, which applies a similar methodology using the FES 2014 tidal model for Africa.

What drew me in was that DE Africa was not just serving raw imagery, but a carefully processed coastline dataset that already accounts for tides and long-term change. This meant I could focus more on the vulnerability modelling itself, instead of spending all my time building the shoreline extraction pipeline from scratch.

What was the focus of your Bachelor’s thesis, and how did DE Africa’s data and services support your research?

My thesis, “Assessment of Coastal Vulnerability to Sea Level Rise: A Case Study of the Kenyan Coastline,” developed a multi-criteria Coastal Vulnerability Index (CVI) for Kenya. I created separate indices for physical exposure and socioeconomic sensitivity, then combined them to map overall vulnerability to sea level rise along the coast.

DE Africa underpinned the shoreline-change component of this work. I used DE Africa’s Landsat 7, 8 and 9 surface reflectance archive (2000–2024) together with the FES 2014 tidal model to derive shoreline positions at four-year intervals. These analysis-ready datasets allowed me to extract median waterline positions under comparable tidal conditions, export them as GeoTIFFs, convert them to vectors in ArcGIS, and then run the Digital Shoreline Analysis System (DSAS) to compute long-term rates of erosion and accretion using methods such as Weighted Linear Regression.

Without the DE Africa Coastlines methodology and underlying datasets, it would have been extremely difficult to generate a consistent 24-year shoreline history for the entire Kenyan coast.

Which tools or datasets on DE Africa were most useful in your work, and why?

The key DE Africa resources in my thesis were:

• Landsat 7/8/9 Surface Reflectance (2000–2024) – These analysis-ready time series formed the backbone of my shoreline extraction workflow. Their consistency and temporal depth were ideal for detecting subtle erosion and accretion trends over two decades.
• FES 2014 Tidal Model as implemented in DE Africa Coastlines – Tide-tagged observations allowed me to filter scenes to a consistent tidal window, approximating 0 m sea level. This greatly reduced noise from tidal variability and made the derived shorelines comparable through time.
• DE Africa Sentinel-2 10 m composite imagery – I used this high-resolution composite as a base for detailed visual assessment and to support the derivation of coastal geomorphology classes that fed into the physical vulnerability indicators.
  

Can you share one achievement or insight from your thesis that DE Africa helped make possible?

A major achievement was producing a quantitative map of shoreline change for the entire Kenyan coastline from 2000 to 2024, and integrating those rates directly into the Coastal Vulnerability Index.

Using DE Africa’s tide-normalised Landsat archive, I generated shorelines at four-year intervals and used DSAS with Weighted Linear Regression to estimate long-term erosion and accretion trends. This revealed hotspot segments where sustained shoreline retreat coincides with densely populated or economically important areas. In some locations, the erosion signal was only visible once noisy tidal effects were removed, which would not have been possible without the FES 2014-based coastline processing used by DE Africa.

This insight strengthened the final vulnerability maps and provided a clearer evidence base for where adaptation efforts should be prioritised.

What’s the biggest challenge you see for EO adoption in African research, and do you have any ideas to overcome it?

One of the biggest challenges is the mismatch between decision-making needs (which are often very local) and the spatial resolution or thematic detail of many freely available EO products. For example, coastal planners may need to understand vulnerability at the scale of specific beaches, infrastructure corridors, or neighbourhoods, but much of the open data remains relatively coarse or lacks detailed geomorphological and geological context.

To overcome this, I see a strong need to expand continent-wide EO products with higher spatial resolution and richer thematic detail, especially for coastal zones, combine satellite-derived coastlines with country-specific geomorphology, geology, and land-use layers, so that vulnerability analyses can reflect local conditions more accurately, and encourage collaborations where national agencies, universities, and platforms like DE Africa jointly validate and refine these products with local field data, so they become trusted references for planning.

What would you like to see DE Africa achieve next for researchers across the continent?

• Develop higher-resolution coastline and nearshore products, including country-specific shoreline segments that resolve smaller bays, estuaries, and reef-lined coasts important for local communities.
• Provide dedicated coastal geomorphology and coastal geology layers for Africa, derived from Sentinel-2 and other sensors and validated with local expertise. This would be incredibly valuable for CVI-style studies, hazard mapping, and ecosystem management.
• Offer machine-learning-ready services in the DE Africa Sandbox, such as example notebooks and pipelines for coastal classification, change detection, and risk modelling, so researchers can build and share reproducible GeoAI workflows on top of DE Africa datasets.

One piece of advice for a young researcher wanting to use DE Africa’s tools?

At first, working with the Sandbox and all the new models and datasets felt daunting for me, but once I allowed myself to start, stop, restart, and learn at my own pace, everything became much easier. So my advice is simple: just start. Pick one small task and explore it without pressure. And make full use of the trainings. Chances are that if you want to use something on DE Africa, there is already specialized training for it, and it will make the whole process much smoother.