11-18, 13:30–16:30 (Pacific/Auckland), WF502
The workshop will include a) field survey component where participants will be able to walk around Auckland to collect data points in QField for QGIS and b) a computer lab component where participants will use Digital Earth Pacific Python Notebooks to generate a land cover map.
FOSS4G abstracts:
Workshop abstract:
Participatory mapping field survey and computer lab: QField integration into machine learning landcover classification within Digital Earth Pacific.
3-hour workshop
The objective of the workshop is to share a workflow to allow for field data, calibration, and validation of a land cover classification model output.
Land cover classification:
Land cover classification forms serves numerous functions including land cover accounting, monitoring land use changes, biodiversity and conservation monitoring, measurements of urban and agricultural expansion as well as forest inventories and national greenhouse gas inventories.
This workshop will include two main components:
1. a field survey component where participants will be able to walk around Auckland to collect data points in QField for QGIS.
2. a computer lab component where participants will use Digital Earth Pacific Python Notebooks to generate a land cover map based on the data collected in their respective surveys.
These participatory mapping workflows enable users from a range of disciplinary backgrounds to contribute to land cover mapping outputs. These outputs may be used for a range of applications including land cover classification.
The learning outcomes of this workshop will include the following:
1. Participants will learn how to collect field data using QField
2. Participants will also be able to ingest this data into Digital Earth Pacific and through a Jupyter Notebook Environment
3. Participants will be able to build on introductory levels of Python programming knowledge.
Within QField and Python, participants will be making use of the following tools and libraries:
QField workflows to be covered:
Point data collection Collection of points for different land cover classes
Transects Transects
Polygons Collecting polygon areas of interest
Accuracy assessments Ensuring data collected is within set thresholds of horizontal accuracy
Python libraries to be covered:
Pandas / Geopandas Vector data analysis and plotting
odc-geo Web map plotting
Rasterio Raster data analysis and plotting
odc.stac Loading satellite data through Digital Earth Pacific Spatiotemporal Asset Catalogues.
Nick is currently completing his PhD through a joint cotutelle program between the University of the South Pacific and the Australian National University. His research focuses on ridge-to-reef environmental monitoring as well as GIS environmental modelling and remote-sensing land-sea frameworks through riparian corridors. He completed his MSc in the water science specialisation through courses in both the Fenner School as well as the Research School of Earth Sciences at the ANU. Nick also completed his MSc thesis research on quantifying the impacts of in-river gravel extraction on sediment transport in Fiji.
Nick's research areas and skills include: GIS and remote sensing, hydrological and environmental modelling, python, FullCAM carbon accounting, field sampling and measurements of surfacewater and groundwater chemical, geophysical and hydrological parameters and some ecological fieldwork sampling experience forestry biomass carbon assessments as well as sampling of benthic invertebrates and ichthyofauna.
He has worked in a range of Government Departments including the Federal Departments of Agriculture, Water and Environment, the Climate Change Division of the Department of Environment and Energy and the Australian Trade Commission. During this time, Nick also worked in environmental monitoring of the impacts of the Ranger Uranium Mine on the Magela floodplains and creeks adjacent close to Jabiru and Kakadu in the Northern Territory. Nick led a team of volunteers to secure second place in the MAXAR Spatial Challenge regional category through a project that combined Digital Globe sub-metre high resolution imagery with FullCAM modelling to assess regeneration of biomass carbon in the context of the 2019-20 bushfire recovery through a case study in Cann River, Gippsland. Nick was also the team lead for the Yadrava na Vanua team that gained first place in the Space for Planet Earth Competition to use satellite data to estimate carbon sequestration. The team was led by students and staff from the University of the South Pacific, University of Fiji and Fiji National University.