SEGH Articles

Aquaculture: Pathway to food security in Kenya

16 July 2018
Dr Andrew Marriott (Centre for Environmental Geochemistry, BGS) and Prof Odipo Osano (University of Eldoret) tell us about their recent voyage on Lake Victoria as part of an exciting project looking into aquaculture and food security in Kenya

In late December we were successful in obtaining funding through a Newton International Links bid with the British Council. This project aims to investigate problems of food insecurity and the implications of anthropogenic pollutants for food safety, the impact on the health and nutritional quality of aquaculture farmed fish, and the potential of fisheries aquaculture to boost food security and sustainability in Lake Victoria, Kenya. Having been successful in the bid, I was looking forward to working with fish again and using the knowledge gained from my years as a Marine Biologist at Bangor University. The project brings together my experiences working on fisheries aquaculture and from knowledge gained working within the Inorganic Geochemistry team in exploring micronutrients, pollution pathways and the associated problems of food security and nutrition in Aquaculture fish. This collaborative project involves the Inorganic Geochemistry team (IG) at BGS with our old friend Prof Odipo Osano from the School of Environmental Sciences at the University of Eldoret (UoE) with whom we have worked closely for over 2 years. This grant funded project also collaborates with Dr Tracey Coffey and Dr Sharon Egan from the School of Veterinary Science (University of Nottingham), and the experienced research team led by Dr Christopher Mulanda Aura at the Kenyan Marine Fisheries Research Institution (KMFRI) based in Kisumu, Kenya.  

 stakeholder discussions

Stakeholder discussions at KMFRI headquarters between the project team and aquaculture producers


Stakeholder discussions were hosted at KMFRI headquarters between the project team and aquaculture producers explaining how the project will address questions related to pollution pathways (Food Safety-human and ecological health), through studying problems associated with sedimentation rates in water from erosion, or changes in land-use and water quality, as well as the effects of toxic metals as drivers of antimicrobial resistance (AMR). Expanding on this, and linking to other BGS activities ( and interests of the Ministry for Agriculture, Livestock and Fisheries, data will be generated to understand the potential for aquaculture to address ‘hidden hunger’ (deficiencies of essential micronutrients for human and animal health). Questions came quick and fast from those attending, with the fisheries-licence holders and the cage-culture representative being extremely interested in what we were intending to do. However, transparency and open talks through public engagements such as this alleviated any concerns they may have had, and their subsequent assistance in locating sites and communicating with local cage managers proved invaluable.


 Team and Stakeholders

The team and the stakeholders


Attention turned to the fieldwork, which was based on KMFRIs main research vessel R.V. Uvumbuzi (Discovery) based on the Winam Gulf in Kisumu. This allowed us direct access and quicker travel between sampling sites within the Winam Gulf to view and sample the intensity of aquaculture cages along the shores of Lake Victoria (Kenyan side).

 aquaculture cages

Aquaculture cages along the shores of Lake Victoria (Kenyan side)


Cage sizes ranged from a few meters (2x2) to the medium 6x4 meters up to the larger 20x20 meters and housing fish from a few thousand to over 20,000 for the larger 20x20 m cages. Most of the fish (Tilapia) are sold at local markets after growing for 6 months (plate size) or in the case of the larger cages (>10m) sold to merchants after growing for up to 8 months for re-sale as fillets. Involving the representative for the aquaculture operators in the stakeholder meeting hosted by KMFRI, provided invaluable as they were able to help in the design of sampling locations, and assist in gaining the trust and co-operation of the managers of each cage site. With the local manager informed in advance of our visit, sampling fish from both within the cages, and from the wild went smoothly and efficiently. Sampling involved all the experience of the project team. Water parameters such as pH, temperature, conductivity and dissolved oxygen were measured either directly from the vessel or from the small canoe brought along for closer inshore work.


The KMFRI team on the canoe


After measurements of the water were obtained, the process of collecting water samples from just below the surface (approx. 1 m), and from depth, was completed using a niskin water sampler system. The collected water was then filtered using 0.45 um filters. Water samples were observed to be slightly discoloured around the shores of Winam Gulf which isa possible indication of high sedimentation rates from land erosion. This colouration cleared when we moved further out to the main channel at Rusinga in the Lake. Finally the collection of a sediment sample was taken using the Van Veen Grab Sampler for both biogeochemical analysis in the UK, and benthic microbiological analysis by the team at KMFRI.

 Van Veen Grab Sampler

The Van Veen Grab Sampler


Most of the sediment collected near the Winam Gulf area near the shore had a strong H2S aroma indicating possible eutrophic conditions in that area. Fish were sampled both from the cages and also where possible from the wild, using fish locally caught by the local fishermen.

 Sampled caged Talapia

Sampled caged Tilapia fish


Fish were processed on board the R.V. Uvumbuzi, and measurements were taken to record the biological data. Finally sub-samples of muscle tissue were taken and vacuum sealed for subsequent analysis of trace metals, and tissue for DNA extraction and Anti-Microbial Resistance studies.


 Marriott fish sampling

Andy Marriott sub-sampling muscle tissue


After 3 long days working on the Lake, the water, sediment and fish tissue samples were finally off-loaded from the RV Uvumbuzi and transported to the School of Environmental Sciences (University of Eldoret). Here sediment samples were partially dried and excess water removed for transportation back to the Inorganic Geochemistry labs in the UK for ICP- MS analysis. Similarly, fish tissue will be measured for trace metals and DNA and AMR analyses at the School of Veterinary Science (University of Nottingham).

Water, fish tissue and sediment geochemistry data in addition to ongoing studies by UoE and KMFRI will help to identify possible pollution pathways from anthropogenic activities, and sediment loading from erosion or changes in land-use for design of further investigation in the Winam Gulf and shores of Lake Victoria. Bio-geochemistry data from this and subsequent investigations will be invaluable to KMFRI in advising on the future siting of aquaculture cages to ensure food safety for human and ecological health. Given that ‘wild’ fish catches are unsustainable in Lake Victoria, aquaculture/farming of fish through sustainable methods is an important approach to food sufficiency and in addressing ‘hidden hunger’ along the shoreline of Lake Victoria and further inland.


I would like to thank the Newton International Links programme and the British Council for funding the project. I would also like to thank the BGS-ODA programme and Center for Environmental Geochemistry for financial support and technical expertise.

Keep up to date

Submit Content

Members can keep in touch with their colleagues through short news and events articles of interest to the SEGH community.

Science in the News

Latest on-line papers from the SEGH journal: Environmental Geochemistry and Health

  • Review: mine tailings in an African tropical environment—mechanisms for the bioavailability of heavy metals in soils 2019-05-27


    Heavy metals are of environmental significance due to their effect on human health and the ecosystem. One of the major exposure pathways of Heavy metals for humans is through food crops. It is postulated in the literature that when crops are grown in soils which have excessive concentrations of heavy metals, they may absorb elevated levels of these elements thereby endangering consumers. However, due to land scarcity, especially in urban areas of Africa, potentially contaminated land around industrial dumps such as tailings is cultivated with food crops. The lack of regulation for land-usage on or near to mine tailings has not helped this situation. Moreover, most countries in tropical Africa have not defined guideline values for heavy metals in soils for various land uses, and even where such limits exist, they are based on total soil concentrations. However, the risk of uptake of heavy metals by crops or any soil organisms is determined by the bioavailable portion and not the total soil concentration. Therefore, defining bioavailable levels of heavy metals becomes very important in HM risk assessment, but methods used must be specific for particular soil types depending on the dominant sorption phases. Geochemical speciation modelling has proved to be a valuable tool in risk assessment of heavy metal-contaminated soils. Among the notable ones is WHAM (Windermere Humic Aqueous Model). But just like most other geochemical models, it was developed and adapted on temperate soils, and because major controlling variables in soils such as SOM, temperature, redox potential and mineralogy differ between temperate and tropical soils, its predictions on tropical soils may be poor. Validation and adaptation of such models for tropical soils are thus imperative before such they can be used. The latest versions (VI and VII) of WHAM are among the few that consider binding to all major binding phases. WHAM VI and VII are assemblages of three sub-models which describe binding to organic matter, (hydr)oxides of Fe, Al and Mn and clays. They predict free ion concentration, total dissolved ion concentration and organic and inorganic metal ion complexes, in soils, which are all important components for bioavailability and leaching to groundwater ways. Both WHAM VI and VII have been applied in a good number of soils studies with reported promising results. However, all these studies have been on temperate soils and have not been tried on any typical tropical soils. Nonetheless, since WHAM VII considers binding to all major binding phases, including those which are dominant in tropical soils, it would be a valuable tool in risk assessment of heavy metals in tropical soils. A discussion of the contamination of soils with heavy metals, their subsequent bioavailability to crops that are grown in these soils and the methods used to determine various bioavailable phases of heavy metals are presented in this review, with an emphasis on prospective modelling techniques for tropical soils.

  • Pesticides in the typical agricultural groundwater in Songnen plain, northeast China: occurrence, spatial distribution and health risks 2019-05-25


    Songnen plain is an important commodity grain base of China, and this is the first study on the comprehensive detection of multiple pesticides in groundwater. Based on an analytical method of 56 pesticides, 30 groundwater samples were collected and analyzed. At least 4 pesticides were detected in each sample and 32 out of 56 pesticides were detected. The average detected levels of individual pesticides were approximately 10–100 ng/L. Organophosphorus pesticides and carbamate pesticides were the dominant pesticides, and their percentage of total pesticide concentrations were 35.9% and 55.5%, respectively. Based on the spatial distribution, the characteristic of nonpoint source pollution was indicated in the whole study area except for a point source pollution with the influence of a sewage oxidation pond. Nine core pesticides and three distinct clusters of the core pesticides with various concentration patterns were revealed by cluster analysis. Linear regression identified a significant relationship between the cumulative detections and the cumulative concentrations, providing access to identify the outlying contaminant events that deviate substantially from the linear trend. A new insight for prediction of pesticide occurrence was provided by the Pearson correlation between some individual pesticide concentrations and the cumulative detections or the cumulative concentrations. According to health risk assessment, the residual pesticides posed medium risks for children and infants and approximately 90% of risks were composed of β-HCH, dimethoate, ethyl-p-nitrophenyl phenylphosphonothioate and methyl parathion. These findings contributed to establishing a database for future monitoring and control of pesticides in agricultural areas.

  • Correction to: Potential CO 2 intrusion in near-surface environments: a review of current research approaches to geochemical processes 2019-05-22

    In the original publication of the article, the third author name has been misspelt. The correct name is given in this correction. The original version of this article was revised.