Floods and drought are one of the most recurring and devastating natural hazards threatening life and economy. Early warning of the likely occurrence of flood and drought disaster risks and impacts could assist in decision making to support proper disaster risk responses, management and formulation of informed Climate Change adaptation and mitigation strategies for enhanced resilience and disaster risk reductions. Globally and across Africa various initiatives and innovations of multi-hazard early warning systems based on web and mobile information platforms have been developed, however, none fit in the Tanzanian and East African environment. Developing a localized multi-hazard early warning for climate related hazards is essential in reducing loss of life and damage to property. This study aimed at developing an innovative multi-hazard early warning Web based Geographical Information System (GIS) App for flood and drought disaster risks in East Africa, Case of Tanzania.

To develop the Web -Based GIS App, the guiding research questions are centred on user requirement elicitation, identification of vulnerable communities, development of the Web-based GIS App conceptual model and the Web-Based GIS App. For the Web-based GIS App requirement elicitation a thorough document review and stakeholders’ workshop was used. To identify and map the floods and drought vulnerable areas we used a Height Above Nearest Drainage (HAND) model and Vegetational condition index (VCI) were utilized. Thereafter, we used the Palmer Drought Severity Index (PDSI) and deep learning neural networks based on geospatial weather data using convolution long-short-term memory (ConvLSTM) model to predict drought in Dodoma. The Web based GIS App was designed and developed using the RAD methodology in Microsoft .NET framework.

From analysis of flood prone areas indicated Magomeni, Kigogo, Mchikichini, Mburahati, Mabibo, Ndugumbi, Kijitonyama Makumbusho, Mikocheni, Kipawa, kiwalani, Kawe, Kunduchi, Mbweni, Chanika, Pembamnazi, Ksarawe II and Mjimwema, Somangila, Hananasifu, Tandale and Msasani wards as areas which are highly prone to flooding. The results were validated using observed flood points in Tandale wards and flood zones which were identified and mapped using the participatory approach by vulnerable community members in Hanansifu and Msasani Wards. On the other hand, the analysis of drought conditions in Dodoma region depicts Chamwino, Bahi and Central Dodoma to be highly prone to drought risks. Thereafter, the Web-based GIS conceptual model was developed followed by development of the Web-based GIS for dissemination of climate related early warning information for flood and drought disaster risks. We recommend integration of GIS and Early Warning Tools into Existing Policies, Establish Monitoring and Evaluation Frameworks and further Improvement of the Developed App for enhanced Disaster Risk reductions.