Ask a civil engineer, a hydrologist, a county planner, and a solar developer what data they need before their project can proceed, and all four will give you the same answer: an elevation model. The form it takes and the accuracy it must achieve will differ entirely between them — but all four are asking for essentially the same thing: a systematic numerical description of the shape of the ground. A Digital Elevation Model is that description. This article explains what a DEM is, how DEM, DTM, and DSM differ, what sources are available for Kenya and East Africa at no cost, and how to determine whether your project needs a freely available global dataset or a purpose-built survey.

What a Digital Elevation Model Actually Is

A Digital Elevation Model is a raster dataset — a grid of cells covering a geographic area — where each cell contains a single numerical value representing the height of the ground (or surface) at that location, measured above a defined vertical datum. In Kenya, the standard vertical datum is mean sea level (MSL), referenced through the national levelling network maintained by Survey of Kenya. Each cell in the grid represents a specific area on the ground — 30 metres × 30 metres for a Shuttle Radar Topography Mission (SRTM) dataset, 10 metres × 10 metres for Copernicus DEM GLO-30, and as fine as 0.1 metres × 0.1 metres for a high-density drone-derived model of a small engineering site.

The fundamental product of a DEM is a terrain surface from which an almost unlimited range of analytical derivatives can be computed: slope gradients, aspect (which direction a slope faces), hillshade (simulated illumination for visualisation), flow direction (which way water runs downhill), flow accumulation (which cells collect water from upstream), watershed boundaries, viewsheds (what is visible from any point), and curvature (whether the terrain is concave, convex, or planar). These derivatives are the actual inputs that engineers and analysts use — the raw DEM is the source from which they are all derived.

A DEM is stored as a GeoTIFF file or occasionally as an ASCII grid, each pixel carrying a single floating-point elevation value. Opening a DEM in QGIS or ArcGIS and applying an elevation colour ramp is the most common first step — rendering the terrain in shaded relief immediately reveals the landscape's structure: ridgelines, valleys, flat plains, escarpments, and drainage networks. This visual representation, while not the end product, is one of the most powerful orientation tools available to a project team beginning work in unfamiliar terrain.

📐 Vertical Datum in Kenya

All elevation values in a properly referenced Kenyan DEM should be referenced to mean sea level (MSL) via the national levelling network. The national datum is maintained by Survey of Kenya through a network of benchmark monuments. When using freely available global DEMs (SRTM, Copernicus DEM), the elevation values are referenced to the EGM96 or EGM2008 geoid model — an approximation to MSL that may differ by ±1–5 m from the true MSL in Kenya, depending on location. For engineering applications where precise RL values are needed, the DEM should be calibrated against known national benchmark values before use.

DEM vs. DTM vs. DSM: The Difference That Changes Everything

The three terms — DEM, DTM, and DSM — are frequently used interchangeably in project briefs and procurement documents, and this imprecision causes a significant proportion of the mismatches between what a project orders and what it actually needs. The distinctions are not semantic; they reflect fundamentally different things being measured, with real consequences for the analyses they support.

DEM
Digital Elevation Model
The generic term for any gridded elevation dataset. In common usage, DEM is often used interchangeably with DTM — particularly for satellite-derived global products like SRTM and Copernicus, which are technically bare-earth models in many areas but contaminated by vegetation and buildings in others. When a project brief says "DEM" without further specification, clarify whether they mean a bare-earth surface (DTM) or a first-surface model (DSM). The ambiguity matters enormously for forested terrain.
Colloquial use: Any gridded elevation raster Technical use: Often synonymous with DTM Common sources: SRTM, ALOS, Copernicus DEM Kenya issue: Global DEMs include vegetation in forested highlands
DTM
Digital Terrain Model
The bare-earth surface — vegetation, buildings, and structures removed, representing the ground itself. For engineering design (road levels, drainage design, earthworks volumes), flood modelling (water flows over the ground, not over the treetops), and watershed analysis, only a DTM gives valid results. From satellite radar data, a true DTM is difficult to produce in vegetated areas — radar bounces off the canopy. From drone photogrammetry, the DTM requires point cloud classification to separate ground from vegetation. From LiDAR, the multi-return capability makes bare-earth DTM extraction possible even through dense forest.
What it represents: Bare earth — no buildings or trees Required for: Road design, flood modelling, drainage, earthworks Best sources: LiDAR (through vegetation) · Drone in open terrain Do NOT use: DSM or uncorrected DEM for engineering design
DSM
Digital Surface Model
The top of everything — the first surface the sensor encounters, whether that is the ground, a building roof, or the top of a tree canopy. A DSM from drone photogrammetry maps the treetops, not the ground beneath. Most satellite-derived global elevation products are actually DSMs in forested areas — the radar or stereo imagery captures vegetation canopy elevation, not bare earth. DSMs are the correct product for viewshed analysis, solar irradiance mapping, building height extraction, and any analysis where the above-ground surface matters. For drainage design over forested terrain, a DSM is the wrong input and will produce systematically wrong results.
What it represents: Top of everything — canopy, roofs, structures Required for: Solar mapping, viewshed, building heights, 3D city models Best sources: Drone photogrammetry · Sentinel-2 stereo · LiDAR DSM Warning: Most "DEMs" in forested Kenya are actually DSMs
⚠️ The Most Common DEM Mistake in Kenya

Using SRTM 30m data as a DTM for flood modelling or drainage design in the Mau Forest, Aberdares, Mt Kenya, or any highland forested area is a systematic error. In forested terrain, SRTM's radar signal reflects off the canopy — producing elevation values that are 5–20 m higher than the actual ground. A flood model built on this "DEM" simulates water flowing over the canopy surface, not the actual ground — producing a drainage network that is wrong in ways that cannot be corrected by simply applying a fixed offset. If your project is in forested terrain, you need a LiDAR-derived DTM or a field-validated model — not SRTM.

Free DEM Sources Available for Kenya

For regional-scale analysis, pre-feasibility work, and projects where ±5–15 m vertical accuracy is acceptable, several high-quality DEMs are available at no cost and can be downloaded for any part of Kenya within minutes. Understanding what each dataset is, how it was produced, and where its limitations are most significant prevents the misapplication that undermines analysis.

Dataset Resolution Vertical Accuracy Type Best Use in Kenya Cost
SRTM 1 Arc-Second (30m) ~30 m ±5–10 m (forested ±15–20 m) DSM in forests · Near-DTM in open areas Watershed delineation, regional slope analysis, prefeasibility corridor selection Free — USGS EarthExplorer
Copernicus DEM GLO-30 ~30 m ±4 m (open) · ±8 m (forested) DSM — WorldDEM derived Better accuracy than SRTM in open terrain · National-scale slope, aspect, contours Free — Copernicus Open Access Hub
ALOS World 3D (AW3D30) ~30 m ±5 m globally · ±3 m open DSM — optical stereo Good for open highlands · Available through OpenTopography or JAXA portal Free — JAXA / OpenTopography
Copernicus DEM GLO-10 ~10 m ±2–4 m (open terrain) DSM — TanDEM-X derived Prefeasibility road design, county planning, urban slope analysis Free — AWS Open Data · Copernicus
NASADEM ~30 m ±4–8 m Reprocessed SRTM — improved voids Better void filling than SRTM over Mt Kenya and highland areas Free — NASA EarthData
ICESat-2 Track Elevations Track-only (sparse) ±3–4 cm (track points) Point elevations — not gridded Calibrating/validating other DEMs · Not suitable as primary DEM Free — NASA Earthdata
Drone Photogrammetry DTM 0.05–2 m ±3–15 cm with GCPs DTM (open terrain) · DSM (vegetated) Engineering design, earthworks volumes, detailed site topography Survey fee — Geopin
UAV-LiDAR DTM 0.1–1 m ±5–15 cm through vegetation True DTM — through canopy Road design in forest, dam sites, flood modelling in vegetated catchments Survey fee — higher than photogrammetry

Six Applications Where the DEM Choice Determines the Answer

🌊
Flood Modelling and Drainage Design
HEC-RAS · HEC-HMS · SWAT · ±30 cm DTM
Flood modelling software (HEC-RAS for 2D floodplain modelling, HEC-HMS for rainfall-runoff) is acutely sensitive to the quality of the terrain input. The model routes water downhill from cell to cell — if the DEM is wrong by 2 m (easily within SRTM error range), the simulated flood extent shifts by hundreds of metres in flat floodplains. County engineers sizing culverts for road crossings, NEMA requiring floodplain delineation for EIA, and structural engineers designing flood-resistant foundations all need a validated DTM — not the free SRTM data that represents canopy elevation in highland Kenya. In the Nairobi basin, the Athi River floodplain, and the low-lying areas around Lake Victoria, SRTM errors of 3–8 m translate directly to wrong culvert sizes, misidentified flood zones, and building permits issued on land that will flood.
Minimum DEM: 10 m Copernicus GLO-10 for prefeasibility Engineering grade: 1–2 m DTM from drone/LiDAR survey Critical issue: Open terrain only for photogrammetry Kenya example: Nairobi basin, Athi River, Yala Swamp
🛣️
Road and Infrastructure Alignment Design
Corridor Selection · Earthworks · KeNHA · County Roads
Road design uses the DEM in two fundamentally different ways: for corridor selection (identifying the optimal alignment through a landscape by analysing slope, curvature, and cross-country obstacles at 30–100 m resolution) and for detailed vertical alignment and earthworks design (requiring cross-section profiles at 20–50 m intervals to ±15 cm accuracy). The first task is appropriate for SRTM or Copernicus DEM GLO-10 in open terrain. The second requires a field survey — either drone photogrammetry or LiDAR — at the design resolution the engineer needs. The most common planning failure in Kenyan road projects is using the free 30 m dataset for detailed design earthworks estimation, producing cut/fill volume estimates that are wrong by 20–40% and triggering variation orders on construction contracts.
Corridor selection: SRTM / Copernicus 30m acceptable Detailed design: 1–2 m drone DTM or LiDAR required Earthworks accuracy: ±5% from 1m DTM vs ±30% from SRTM Kenya example: KeNHA A2–A8 routes, county road upgrades
💧
Watershed and Catchment Delineation
WRMA · Dam Design · Irrigation · Hydrology
Watershed delineation — identifying the land area that drains to a specific point — is one of the most common DEM-based analyses in Kenya, required for dam design, water resource management, irrigation scheme planning, and WRMA licensing applications. The delineation algorithm fills DEM depressions and routes flow downhill from every cell to its pour point. The accuracy of the resulting watershed boundary depends on the DEM's terrain representation: in flat terrain, a 5 m elevation error can shift the watershed boundary by kilometres. For catchment areas upstream of dams — where the calculated area determines the design flood discharge, and the design flood determines the spillway size — DEM accuracy is a safety parameter, not an analytical nicety.
Prefeasibility: SRTM acceptable for large (>100 km²) catchments Dam design: Field-validated 10m DTM minimum Tools: ArcHydro, QGIS SAGA, Google Earth Engine Kenya example: Tana, Athi, Nzoia, Nyando catchments
☀️
Solar Potential Mapping
PVGIS · Slope/Aspect · Shadow Analysis · Off-Grid
Solar photovoltaic and solar water heating project siting requires terrain analysis to identify slopes facing approximately north in the southern hemisphere (optimum for fixed-tilt panels), calculate shading from surrounding terrain and vegetation at different times of day and year, and determine access and infrastructure logistics. For national or county-scale solar potential mapping, the free Copernicus DEM GLO-30 provides adequate slope and aspect data. For site-specific design — particularly for smaller off-grid systems in hilly terrain — the shading analysis sensitivity means a 5 m slope error on a nearby ridge can incorrectly classify a slope as shaded when it receives 4 hours of daily direct radiation. Solar developers in the Rift Valley and highland areas should be cautious about using 30 m SRTM for site-specific shadow analysis.
National mapping: SRTM / Copernicus 30m adequate Site analysis: 10m or better preferred Tools: PVGIS (EU), r.sun (GRASS/QGIS), ArcGIS Solar Analyst Kenya example: Rift Valley off-grid, Northern Frontier solar farms
⛰️
Landslide Hazard and Slope Stability Mapping
NEMA EIA · Road Safety · Settlement Siting
Kenya's highland counties — Kericho, Nandi, Elgeyo-Marakwet, Murang'a, Nyamira — experience periodic and sometimes catastrophic landslides, particularly on steep tea-growing slopes following prolonged rainfall. DEM-derived slope maps are the first layer in any landslide susceptibility assessment, identifying terrain above critical slope angles (typically 25–35° for Kenya's soil and geology types). Combined with geology, land cover, and rainfall data, a slope map from a 10–30 m DEM provides the spatial framework for hazard zonation that informs settlement siting, road routing, and NEMA EIA requirements. The 2023 Kericho and 2024 Kenol-Makuyu landslide events both occurred on slopes that would have been classified as high-hazard in any competent DEM-based assessment.
Hazard screening: 30m Copernicus adequate for county scale Detailed assessment: 5–10m DEM preferred for specific sites Tools: QGIS slope tool, SAGA Terrain Analysis, ArcGIS Kenya priority areas: Western and Central Highland escarpments
🌾
Irrigation Scheme Design and Precision Agriculture
NIFC · Gravity Irrigation · Slope Mapping · Drainage
The National Irrigation and Drainage Authority (NIDA, formerly NIB) and county governments planning irrigation schemes need terrain data to design gravity canal alignments, assess field drainage potential, calculate earthworks for terrace construction, and map areas within reach of irrigation infrastructure. For large-scale scheme design (>500 ha), a 10–30 m DEM provides the framework for canal siting and water delivery radius estimation. For field-level irrigation design — where canal grades must be accurate to 0.1% and field levelling design requires ±5 cm accuracy — a field-surveyed DTM from drone photogrammetry or total station ground survey is essential. Precision agriculture applications (variable-rate input application, drainage tile design) additionally require a DEM accurate enough to differentiate micro-topography within individual fields.
Scheme siting: SRTM / Copernicus 30m for prefeasibility Canal design: 1–5m drone DTM required Field levelling: ±5 cm drone survey + GCPs Kenya example: Mwea, Bura, Galana-Kulalu, Lake Naivasha schemes

Choosing the Right DEM for Your Project: A Decision Guide

The question "which DEM should I use?" has no universal answer — it depends on four variables: the application (what analysis you need to run), the terrain type (open, vegetated, or mixed), the required accuracy (what error is tolerable in your outputs), and the project stage (prefeasibility vs. detailed design vs. construction). The guide below maps these variables to a recommended DEM source.

DEM Selection Decision Guide — Kenya Projects
Match your project to the appropriate data source
Project Scenario
Terrain / Stage
Recommended Source
National / county watershed mapping
Prefeasibility, large catchment (>100 km²)
Open highlands, savannah, arid
SRTM 30m or Copernicus GLO-30
Road corridor selection
Alignment options study, multiple routes
Mixed terrain — any cover type
Copernicus GLO-10 (10m)
Flood hazard mapping
County-level flood zone designation
Open floodplain, urban
Copernicus GLO-10 + field validation
Solar potential assessment
Regional or site siting study
Open terrain, highlands
SRTM / Copernicus 30m adequate
Landslide hazard screening
EIA or county development plan
Highland escarpments
Copernicus GLO-10 preferred
Road detailed design
Earthworks volumes, vertical alignment
Open terrain, arid corridor
Drone photogrammetry 0.5–2m
Flood hydraulic modelling
HEC-RAS 2D, engineering-grade
Open floodplain, urban
Drone photogrammetry 0.5m DTM
Dam catchment hydrology
Design flood, spillway sizing
Mixed — open + some forest
Drone DTM + field survey validation
Road design in forested corridor
Earthworks through vegetation
Dense forest — canopy present
UAV-LiDAR DTM (only option)
Irrigation canal design
Field-level grades, earthworks
Agricultural land, open
Drone DTM ± 5 cm + GCPs
Dam site engineering survey
Foundation, reservoir volume
Forested / mixed catchment
UAV-LiDAR DTM + bathymetric survey
Building/site engineering
Foundation levels, earthworks <10 ha
Any — cleared site typical
Total station / drone 0.05–0.5m

How to Download a Free DEM for Your Kenya Project

Accessing the free global DEMs described in this article requires nothing more than an internet connection and a free account registration. The following are the most practical access routes for Kenyan project teams.

For SRTM 30m and NASADEM

Register a free account at NASA EarthData (earthdata.nasa.gov). Search for "SRTM 1 Arc-Second Global" or "NASADEM" in the EarthData Search portal. Navigate to the tile covering your area of interest (Kenya falls in tiles N00E034 through N05E042 approximately, depending on location), select the tile, and download the GeoTIFF. Alternatively, in QGIS, install the SRTM Downloader plugin which allows direct download to your canvas extent. In Google Earth Engine, SRTM is available as ee.Image("USGS/SRTMGL1_003") — ready to use in any GEE script without downloading.

For Copernicus DEM GLO-30 and GLO-10

The Copernicus DEM is available through multiple access points. The simplest for project teams: via OpenTopography (opentopography.org) — draw a rectangle over your area of interest and select "Copernicus Global DEM" from the product list. Download as GeoTIFF, typically within seconds for areas under 10,000 km². For larger areas, access through the AWS Open Data Registry (Copernicus DEM is hosted in the copernicus-dem-30m S3 bucket) or the ESA Copernicus Data Space Ecosystem — a registration is required but access is free.

For ALOS AW3D30

Download from the JAXA portal (www.eorc.jaxa.jp) or through OpenTopography. The 5m resolution product (AW3D) is commercial; the 30m version (AW3D30) is free for non-commercial use after registration.

💡 Quick Start in QGIS

The fastest way to get a DEM for any Kenya location in QGIS: install the SRTM Downloader plugin (Plugins → Manage and Install → search "SRTM Downloader"). Zoom to your area of interest in QGIS, run the plugin, and it downloads the SRTM tiles automatically, merges them, and loads them into your project. For Copernicus DEM, use QuickMapServices or the OpenTopography plugin for direct access. For visualisation, apply a colour ramp (single band pseudocolour in layer properties) and enable Hillshading in the raster layer's symbology to immediately see the terrain structure. Export contours using Raster → Extraction → Contour in QGIS at whatever interval your project requires.

A 30-metre free DEM is the right tool for understanding a landscape. A 1-metre survey-grade DTM is the right tool for designing infrastructure across it. Knowing which you need — and when to upgrade — is what separates a competent GIS analyst from an expensive mistake.
📍 From the Geopin Field · North Horr–Ileret, Marsabit County

The design team for the 115 km A4 Road project in Marsabit County initially approached road vertical alignment using SRTM 30m data — a reasonable starting point for corridor prefeasibility in the arid, open terrain of the Chalbi Desert. SRTM performs well in open, desert terrain: vertical accuracy of ±4–6 m is achievable, and the 30m resolution is adequate for broad alignment options. However, the 14 seasonal wadi crossings along the route required hydraulic analysis for culvert sizing — each wadi's catchment needed to be delineated precisely and the design flood estimated from rainfall data. For the wider shallow wadis, SRTM catchment delineation was adequate as a first estimate. For three deep, narrow gorge crossings where catchment boundaries were sensitive to small DEM errors, Geopin deployed drone photogrammetry at 4 cm/pixel GSD, producing 0.5 m DTMs of each crossing corridor. The refined catchment areas at these three crossings changed the design flood estimates by 15–28% compared to the SRTM-based calculations — a difference that would have resulted in significantly undersized culverts under significant design floods.

DEM and Terrain Survey Services

Survey-Grade DTMs for Engineering and GIS Projects Across Kenya

Geopin produces drone photogrammetry and LiDAR-derived DTMs at the resolution and accuracy your project requires — with GCP-validated accuracy reports certified for engineering design use.

Commission a Terrain Survey →
Tags: Digital Elevation Model DEM DTM DSM Kenya SRTM Kenya Copernicus DEM Flood Modelling Kenya Watershed Analysis Road Design DEM Solar Mapping Kenya LiDAR DTM Drone Photogrammetry DTM GIS Kenya HEC-RAS Kenya
About the Author
GC
Geopin Consult GIS & Remote Sensing Team
GIS Specialists · Nairobi, Kenya

Geopin's GIS and survey team produces terrain models for engineering design, flood analysis, and catchment hydrology across Kenya and East Africa — from 30m free global datasets used in prefeasibility to sub-10 cm UAV-LiDAR DTMs for detailed engineering design. Our deliverables include GCP-validated accuracy reports for all survey-grade terrain models.