Earth Observation for marine resource management

Kenya’s coast holds a network of marine protected areas that has been building since the 1960s: Kiunga in the Lamu archipelago, Malindi and Watamu, Mombasa, and Kisite-Mpunguti in the south. Together they cover coral reef, seagrass and mangrove systems that support both biodiversity and coastal livelihoods.

Designation is a legal act. Management is a continuous one, and it is harder.

The measurement problem

Managing a marine area requires knowing what is happening in it, and marine environments resist that in ways terrestrial ones do not. There is no vegetation to photograph from a road. Boundaries are invisible. Conditions change on a timescale of hours. Access requires a boat and good weather.

The result is that management decisions often rest on infrequent survey data and on reported activity, both of which are sparse relative to the area being managed.

What satellites can contribute

Earth Observation does not replace field survey, and anyone claiming it does has not spent time on a reef. What it provides is regular, consistent, area-wide observation between surveys.

Habitat extent and change. Multispectral imagery supports mapping of mangrove, seagrass and shallow reef, and more importantly supports measuring how their extent changes over time. Mangrove is particularly tractable: it has a distinctive spectral signature, it sits at the land-water boundary where it is easy to isolate, and its loss is both ecologically significant and a carbon question.

Water quality indicators. Turbidity, suspended sediment and chlorophyll proxies can be derived from optical data. These are indicators rather than measurements, and they need local calibration to mean anything quantitative, but as a time series they show where and when conditions are changing.

Coastal change. Shoreline position, erosion and accretion are directly observable and are relevant to infrastructure planning as well as to conservation.

Activity. Radar detection of vessels shows presence in and around managed areas independently of what those vessels report, which is the subject of a separate note.

Time series beats snapshots

The single most useful property of satellite data in this domain is not resolution. It is repetition.

A high-resolution image of a reef on one day says something about that day. A decade of consistent observation says whether the system is degrading, and how fast, and whether an intervention changed the trajectory. Management questions are almost always questions about change, and change requires a baseline.

This is also where the free and open data policy matters most. A consistent archive going back years, acquired on a fixed schedule with stable calibration, is worth more for change detection than a handful of sharper images acquired opportunistically.

Being honest about limits

Optical observation of shallow water is genuinely difficult. The signal reaching the sensor has passed through the atmosphere, then the water column, reflected off the bottom, and returned through both again. Water depth, turbidity and surface state all affect it. Atmospheric correction over water is its own specialist problem.

The result is that satellite-derived habitat maps carry real uncertainty and require validation against field data. Presenting them without that caveat would misrepresent what the method can support.

Where this leads

The useful framing is not satellite versus field survey. It is using each for what it is good at: field survey for calibration, validation and everything requiring physical presence; satellite observation for extent, frequency and consistency.

For institutions managing large marine areas with limited budgets, that combination stretches the field capacity they have, by telling them where it is most needed.