Synthesising multiple observations into annual environmental condition reports: The OzWALD system and Australia's Environment Explorer

Routine, nation-wide environmental data collection, analysis, and reporting remains a challenge in Australia. We wished to test what environmental data is already being collected and can be usefully synthesised, interpreted and reported upon, and to develop a deeper understanding of the scientific and technological opportunities and challenges. To that end, we developed a fully functional annual environmental reporting process, incorporating an automated workflow for data acquisition, integration and summary; a website to deliver the summary the data in a visually informative form; and an annual interpretation, reporting and communication process. Here we report mainly on the technological aspects. The Australian Water and Landscape Dynamics (OzWALD) system is a data production workflow that can be run on demand. It includes the following components: (1) Data acquisition from several gridded climate and satellite data sources; (2) Data reformatting, e.g., spatial sub-setting, reprojection, temporal aggregation, and vector-grid transformation; (3) Data fusion and enhancement, referring to the blending of like data into a single best-estimate data set, and the improvement of desirable data characteristics, respectively. This results in a blended satellite-gauge precipitation data set, an internally-consistent set of dynamic water and land surface properties derived from satellite products (WALDMORF), and downscaled climate data at 500-m resolution; (4) Model-data assimilation, where a biophysical model of the same lineage as the AWRA Community Modelling System is infused with the observational data to estimate additional variables (soil moisture, streamflow generation and vegetation carbon uptake); and (5) Statistical summary by year, region and land use type, deriving temporal and regional statistics, used predominantly for web-based visualisation. A website, Australia's Environment Explorer (AEE, www.ausenv.online), was developed to allow users visualise and explore environmental changes by region, location or land use type. The data can be queried in different ways and are visualized in charts. There have been four annual updates to the AEE since 2016, coinciding with the accompanying Australia's Environment report and public briefing. In addition to the AEE, the report also synthesises information from other sources and interprets specific events and temporal trends in global or non-spatial data. Our experiences demonstrate that it is feasible to produce useful, observation-based annual environmental reports. Developing the experimental system and reporting process produced valuable insights, including: (1) the unstable ad continually evolving spatial environmental data services environment is a challenge for maintaining an operational workflow; (2) the rapid development of open source and cloud technologies provides major opportunities and efficiencies; (3) regular, detailed and accurate land cover and land use mapping will be required to achieve successful environmental accounting; (4) the abundance of past, current and future satellite mission observations provides many opportunities for environmental reporting; and (5) obtaining regular and reliable data on biodiversity remains a major challenge. Our goal is to continue, and as much as possible, improve the annual modelling and reporting process until a similar service is available from another source.