Classification and prediction of river network ephemerality and its relevance for waterborne disease epidemiology

The transmission of waterborne diseases hinges on the interactions between hydrology and ecology of hosts, vectors and parasites, with the long-term absence of water constituting a strict lower bound. However, the link between spatio-temporal patterns of hydrological ephemerality and waterborne disease transmission is poorly understood and difficult to account for. The use of limited biophysical and hydroclimate information from otherwise data scarce regions is therefore needed to characterize, classify, and predict river network ephemerality in a spatially explicit framework. Here, we develop a novel large-scale ephemerality classification and prediction methodology based on monthly discharge data, water and energy availability, and remote-sensing measures of vegetation, that is relevant to epidemiology, and maintains a mechanistic link to catchment hydrologic processes. Specifically, with reference to the context of Burkina Faso in sub-Saharan Africa, we extract a relevant set of catchment covariates that include the aridity index, annual runoff estimation using the Budyko framework, and hysteretical relations between precipitation and vegetation. Five ephemerality classes, from permanent to strongly ephemeral, are defined from the duration of 0-flow periods that also accounts for the sensitivity of river discharge to the long-lasting drought of the 70's-80's in West Africa. Using such classes, a gradient-boosted tree-based prediction yielded three distinct geographic regions of ephemerality. Importantly, we observe a strong epidemiological association between our predictions of hydrologic ephemerality and the known spatial patterns of schistosomiasis, an endemic parasitic waterborne disease in which infection occurs with human-water contact, and requires aquatic snails as an intermediate host. The general nature of our approach and its relevance for predicting the hydrologic controls on schistosomiasis occurrence provides a pathway for the explicit inclusion of hydrologic drivers within epidemiological models of waterborne disease transmission.

Space and time predictions of schistosomiasis snail host population dynamics across hydrologic regimes in Burkina Faso

Theophile Mande, Francisco Javier Perez Saez, Andrea Rinaldo

The ecology of the aquatic snails that serve as obligatory intermediate hosts of human schistosomiasis is driven by climatic and hydrological factors which result in specific spatial patterns of occurrence and abundance. These patterns in turn affect, jointly with other determinants, the geography of the disease and the timing of transmission windows, with direct implications for the success of control and elimination programmes in the endemic countries. We address the spatial distribution of the intermediate hosts and their seasonal population dynamics within a predictive ecohydrological framework developed at the national scale for Burkina Faso, West Africa. The approach blends river network-wide information on hydrological ephemerality which conditions snail habitat suitability together with ensembles of discrete time ecological models forced by remotely sensed estimates of temperature and precipitation. The models were validated against up to four years of monthly snail abundance data. Simulations of model ensembles accounting for the uncertainty in remotely sensed products adequately reproduce observed snail demographic fluctuations observed in the field across habitat types, and produce national scale predictions by accounting for spatial patterns of hydrological conditions in the country. Geospatial estimates of seasonal snail abundance underpin large-scale, spatially explicit predictions of schistosomiasis incidence. This work can therefore contribute to the development of disease control and elimination programmes.

2019

Use of a Distributed Sensor Network to Parameterize a Model of Flows between the Soil, Vegetation, and Atmosphere in a Mixed Savanna-Agricultural Catchment in South Eastern Burkina Faso.

Natalie Claire Ceperley

The hydrologic processes underlying a natural savanna in a semi-arid environment is altered when savanna is converted to agriculture. Inﬁltration is often observed to decrease while rapid runoff increases to the detriment of crop productivity, downstream populations and local water sources. Planting a mix of vegetation in cultivated fields, mimicking the dispersal of trees in a savanna has been proposed as a solution to help buffer these consequences. However the hydrologic behavior of savanna, agricultural, and agroforestry land cover is not fully understood for the Sudanian vegetation zone of West Africa and the consequences of land use changes are debatable. We use a distributed sensor network of seventeen wireless meteorological stations including soil moisture sensors in addition to two eddy covariance stations and sap ﬂow probes dispersed across cultivated rice and millet ﬁelds, natural savanna, fallow ﬁelds, and around agroforestry trees, specifically Sclerocarya birrea, to understand land cover controls on variations in evapotranspiration and infiltration over a semi arid watershed in Southeastern Burkina Faso. N ormalized difference vegetation indexes taken from weekly MODIS images as well as personal field observations were used to inform seasonal and spatial variations in albedo, rate of transfer to ground heat flux, and roughness length. Measurements are used parameterize a distributed model of energy and hydrologic fluxes between the soil and atmosphere controlled by the vegetated surface and presence of agroforestry trees. We use this model to further understanding of the effect of Sudanian land cover on the water and energy balances by exploring the effects of modifications in the arrangement and dispersion of vegetation types. Outreach work in the community is on-going and our hope is that the results of our modeling will inform local farming practices.

2012

A field-based modelling framework of the ecohydrology of schistosomiasis

Francisco Javier Perez Saez

Successful control of schistosomiasis, a water-borne parasitic disease, is challenged by the intricacy of the wormâs lifecycle, which depends on aquatic snail intermediate hosts, and involves environmental, ecologic, and socio-economic factors. Current strategies rely on deworming through mass drug administration which however do not protect against reinfection and the persistence of hotspots. It is recognized that multifaceted approaches will be necessary to reach elimination, whose development will require a renewed focus on the diseaseâs social-ecological drivers. Taking cue from the hydrological underpinning of these drivers, this Thesis aims at developing an ecohydrological approach to schistosomiasis with a view to identifying and exploiting the points in which its cycle can be broken. Schistosomiasis is a poverty-reinforcing disease affecting more than 150 million people in sub-Saharan Africa, being the parasitic disease causing the largest health burden after malaria. However, the impairing morbidity it causes has been undervalued in the past, qualifying it as a neglected tropical disease. Moreover, water resources development often exacerbate transmission, posing scientific and ethical challenges in addressing the ensuing trade-off between economic development and public health. The relevance of this Thesisâ work lies in furthering tools to offset this trade-off by unlocking the predictive appraisal of the social-ecological drivers of transmission. An integration of fieldwork applied in Burkina Faso (West Africa) and theoretical methods are employed to address this aim. This Thesis establishes the use of spatially explicit mathematical models of schistosomiasis at the national-scale, allowing to study the effect of human mobility and spatial heterogeneity of transmission parameters. Weekly ecological samplings of snail abundance and continuous environmental monitoring were preformed at three sites along the countryâs climatic gradient, leveraged through ecological modelling. A novel methodology for the large-scale prediction of river network ephemerality allowed for refined snail species distribution models, and the analysis of the diseaseâs geography in link with socio-economic covariates. Finally, surveys and participatory workshops shed light on local-scale water contact patterns. The obtained results substantiate the stance that hydrology is a first-order control of disease transmission. Stability analysis of the spatially explicit model generated additional insight into the impact of the expansion of suitable snail habitat due to water resources development, highlighting the interplay between local and country-wide effects driven by human mobility. Models of snail ecology revealed key hydrological drivers, and disputed density feedbacks. Uncovered phase shifts between permanent and ephemeral habitats were adequately reproduced at the national scale through model regionalization. Characterization and predictions of hydrological ephemerality improved the estimation of the snailsâ ecological range, mirroring the diseaseâs geography. Finally a national-scale association between ephemerality and disease risk was observed, possibly due to human-water contacts aggregation, as supported by preliminary results at village-level. The future incorporation of these ecohydrological findings into spatially explicit models of schistosomiasis is considered promising for optimizing control strategies and attaining disease elimination.