Giessener Elektronische Bibliothek

GEB - Giessener Elektronische Bibliothek

Modeling hydrological fluxes of tropical mountainous watersheds in Kenya using crowdsourced water level data

Weeser, Björn

Zum Volltext im pdf-Format: Dokument 1.pdf (3.658 KB)

Bitte beziehen Sie sich beim Zitieren dieses Dokumentes immer auf folgende
URN: urn:nbn:de:hebis:26-opus-159072

Freie Schlagwörter (Englisch): citizen science , hydrology , modeling , crowdsourcing , telephone survey
Universität Justus-Liebig-Universit√§t Gie√üen
Institut: Department of Landscape Ecology and Resources Management
Fachgebiet: Agrarwissenschaften und Umweltmanagement
DDC-Sachgruppe: Geowissenschaften
Dokumentart: Dissertation
Sprache: Englisch
Tag der m√ľndlichen Pr√ľfung: 12.02.2021
Erstellungsjahr: 2020
Publikationsdatum: 18.02.2021
Kurzfassung auf Englisch: Climate change and a growing population alter established water usage pathways in Eastern Africa and create an urgent need for effective and sustainable water management strategies. However, required data to develop such strategies is often missing, especially in remote regions. This dissertation examines (1) whether water level data collected by citizens can improve the hydrological database, (2) how this data can be used to establish rainfall-runoff models, and (3) the socio-economic background and motivation of citizens to participate in data collection or reasons that prevent them from continuing.
First, a crowdsourced water level monitoring network was established at thirteen locations within the Sondu-Miriu River basin located in Western Kenya. Interested citizens were invited to record water level data and report these values by sending a simple text message using their cellphone. Over a period of 3.5 years 258 citizens reported 3,480 valid data points. Validation against water level data collected by an automatic radar station at one of the sites revealed high data quality.
In a second step, a conceptual rainfall-runoff model was calibrated on water level data collected by citizens using Spearman-Rank-Coefficients between the simulated discharge and the water levels. Considering a water balance filter derived from measured precipitation and remotely sensed evapotranspiration, the model calibrated on crowdsourced data reached a model efficiency close to values obtained from a benchmark model that was built using automatically measured discharge data (Nash-SutcliffeEfficiency of 0.69 compared to 0.88).
Finally, a telephone survey among the participants in the monitoring project revealed that those who submitted data over a long period were generally between 30 and 50 years old and hold a primary or secondary school diploma. Many participants stated that helping water management and conservation purposes were their primary motivation of involvement. Sensitization meetings were mentioned as the main source of information about the project by long-term participants.
This reserach shows that crowdsourced monitoring approaches are a promising additional tool for water resources management, particularly in ungauged or poorly gauged catchments and under limited financial resources. These findings can be used to support the development for sustainable community-based water monitoring programs.
Lizenz: Veröffentlichungsvertrag für Publikationen ohne Print on Demand