2021-09-29, 16:00–16:30, Aconcagua
In Pacific Island Countries, the environmental resources that support livelihoods are distributed across landscapes in a mix of spatial patterns. Capturing the spatial detail of landscape use is important to inform landscape management that is sensitive to these livelihood dependencies. Using information and communication technologies for development (ICT4D) and agile software development processes, a workflow was developed that comprises open-source geospatial software to map and monitor agricultural landscapes. This workflow was co-developed with the Vava’u branch of the Ministry of Agriculture, Food, Forests, and Fisheries (MAFF) of the Government of Tonga.
The workflow consists of mobile GIS to map farms, web-applications to synchronise and store data, and spatial dashboards for data visualisation and analysis. Mobile geospatial data collection uses QField for intra-farm mapping of cropping practices and digital forms to record farm management attributes. A web application has been developed using Express and Python to support data syncing, automatically generating datasets for reporting on cropping practices and landscape conditions, and for secure data storage. A spatial dashboard, built using Shiny and Leaflet, allows non-GIS experts to easily query and visualise landscape data collected in the field and to use this data in landscape decision making.
This workflow has been used by MAFF for an array of data collection and mapping campaigns. Example uses include: mapping the location of vanilla plantations under sub-optimal management condition; identifying where land was under-utilised or left fallow by farmer groups to spatially target fuel and cash resources to increase land under cultivation; and annual crop monitoring to generate island-wide coverage of intra-farm cropping practices to serve as baseline data to track agricultural change through time.
This talk will discuss the software development process including: the needs assessment to identify and prioritise unmet needs for geospatial data and applications; requirements identification and analysis using use case modelling and rapid prototype development and testing; and refinement and deployment of the workflow for agricultural landscape monitoring on the island group of Vava’u. This talk will also elaborate on the implementations of the workflow, highlight lessons learnt through the development process, and highlight areas for future work and expansion.
This talk will present a workflow developed using open-source geospatial software to map and monitor diverse agricultural landscapes in Pacific Island Countries. This workflow has been developed as a partnership between the Ministry of Agriculture, Food, Forests, and Fisheries (MAFF) of the Government of Tonga and Fijian and Australian universities. The talk will consist of three main sections.
The first section will discuss the information and communications for development (ICT4D) and agile software development process used to design and build the workflow. This process started with a needs assessment to identify and prioritise unmet needs for geospatial data and applications within MAFF. Following the needs assessment, developers, MAFF officials, and researchers engaged in an iterative process of use case modelling, prototype development, and field testing to specify the functional and nonfunctional requirements for the workflow to map and monitor agricultural landscapes.
The second section will introduce the software components. Based on a review of over 100 mobile GIS and data collection apps and field testing, QField was identified as the app most suited for in-the-field farm mapping. In order to synchronise and manage data collected on multiple mobile devices and across mapping campaigns, a cloud-based web app was built using Express and Python. This web app provides support for syncing data collected in the field, secure data storage, and allows analysts to download raw and processed data. Analysts and landscape decision makers can visualise data collected in the field using QGIS or a custom spatial dashboard built using Shiny, Leaflet, and DataTables software. The Shiny dashboard aims to allow non-GIS experts easy access to spatial data collected for landscape monitoring and decision making.
The final section will present how the workflow has been used by field officers on the island groups of Vava’u and Tongatapu for a range of farm and landscape mapping campaigns. Field officers on Vava’u mapped the location and extent of all the vanilla plantations and recorded information about management conditions, grower status, variety, and age of crops. This information was used to identify plantations under sub-optimal management conditions and to target efforts to boost production of an important cash crop. In response to Covid-19 induced food supply chain risks, land allocated to grower groups were mapped to identify fallow fields. Capturing this spatial information informed strategies to ensure agricultural land was well-utilised and to allocate fuel and cash resources to farmers to increase cultivation. Maps and visualisations generated from this data were also used in discussions with local Town Officers to pinpoint challenges to bringing more land under cultivation. At present, this workflow is being used to generate island-wide baseline data of intra-farm cropping practices and farm management. Over 1500 fields have been mapped to-date with this information being used to inform planting decisions and to track agricultural change through time.
Duncan, John (1)
Davies, Kevin (2)
Saipaia, Ahi (3, 4)
Varea, Renata (4)
Vainikolo, Leody (3)
Boruff, Bryan (1)
Bruce, Eleanor (2)
Wales, Nathan (3)
(1) UWA School of Agriculture and Environment, The University of Western Australia, Australia
(2) School of Geosciences, The University of Sydney, Sydney
(3) Ministry of Agriculture, Food, Forests, and Fisheries, The Government of Tonga, The Kingdom of Tonga
(4) Geography, Earth Science and Environment, The University of the South Pacific, Fiji
Use cases & applicationsTopic –
FOSS4G implementations in strategic application domains: land management, crisis/disaster response, smart cities, population mapping, climate change, ocean and marine monitoring, etc.Level –
1 - Principiants. No required specific knowledge is needed.Language of the Presentation –
I’m a geographer and spatial data scientist. I conduct research on, and develop open-source GIS and geospatial technologies for, monitoring environmental resources and analysing human-environment interactions.