ANDS Logo

Project promotion materials:

Project Homepage:

http://www.ccrc.unsw.edu.au/sites/default/files/NARCliM/projects.html

Data collections can be seen on:

https://researchdata.ands.org.au/climate-model-downscaling-data-for-impacts-research

Software is available at:

https://bitbucket.org/a_hotz_unsw/climddir/overview

Programming language(s):

Much of the application will be written in python, with other languages used as necessary. A Domain Specific Language (DSL) may also be employed.

Project Members:

Ian Macadam (i.macadam@unsw.edu.au)

David Fuchs (d.fuchs@unsw.edu.au)

Carlo Piva ( )

Adam Hotz ( )

Patrice Constanza ( )

ANDS Contact:

Mingfang Wu (mingfang.wu@ands.org.au)

Project Status:

Completed

Climate Model Downscaling Data for Impacts Research

University of New South Wales

Project Description

The software system developed by the CliMDDIR project is intended to facilitate research into climate change and its potential impacts. Specifically, the project contributes to bridging a gap between climate modellers, who develop and run computer simulations of the Earth’s climate system, and climate change impacts researchers, who seek to assess the potential impacts of climate change on human and natural systems.

Datasets that describe plausible future climate conditions at a local or regional scale are of use to climate change impacts researchers. However, the Global Climate Models (GCMs) that have been historically used for research into the climate system do not provide data that give a plausible representation of the climate on the fine spatial scales that are relevant to many impacts studies. Fortunately, recent advances in computing and model development have allowed climate modellers to create much higher resolution datasets using Regional Climate Models (RCMs). Unfortunately, RCM-derived datasets are often not widely exposed to the impacts community, contain vast quantities of data that are not directly relevant to impacts studies and are generally stored in a binary format that is convenient for climate modellers to use but is not suitable for direct ingestion into the software tools used by impacts researchers. The extraction and reformatting of impacts-relevant data requires the use of specialist software with which many impacts researchers are unfamiliar. It is important that this problem is overcome if the considerable resources now being invested in regional climate modelling are to be fully leveraged.

The CliMDDIR project has developed a software system that makes impacts-relevant output from RCM simulations available to climate change impacts researchers in data formats that they are used to handling. The system reads binary files of RCM output and exports impacts-relevant data in formats used by the impacts community. Users require no knowledge of specialist software as they interact with the system via a web portal that allows them to extract data that is of relevance to them. The portal also allows users to create data collections that can be exposed to other users and to the broader research community via the the ANDS Research Data Australia website.

The CliMDDIR project was undertaken by a project manager and four software developers employed by the Australian Research Council’s Centre of Excellence for Climate System Science with funding and technical advice from ANDS. However, the project relied on close collaboration between the project team and a number of other project partners. Chief among these was the New South Wales and Australian Capital Territory Regional Climate Model (NARCliM) project. This project is currently using RCMs to develop new, high (~10km) resolution climate projections for New South Wales and the Australian Capital Territory that will provide critical information to manage the impacts of climate change on health, settlements, agriculture, weather extremes and services, such as water and energy supplies. The CliMDDIR service has been implemented in such a way that it can evolve into one of a suite of software tools being put in place to provide access to data derived from RCM simulations performed for the NARCliM project. This has only been possible through close collaboration between the CliMDDIR and NARCliM project teams, and the willingness of the NARCliM team to allow part of the CliMDDIR service to be hosted on its hardware.

To ensure that the CliMDDIR software system met the needs of the impacts community, the project team worked in close collaboration with four groups of climate change impacts researchers:

Agricultural Impacts Group
Research on wheat yields in New South Wales led by Dr De Li Liu (NSW DPI)

Hydrological Impacts Group
Research on rivers and wetlands led by Prof Richard Kingsford (UNSW)

Health Impacts Group
Research on human health led by Dr Hilary Bambrick (UWS)

Ecological Impacts Group
Research on species distributions led by Dr Linda Beaumont (Macquarie University)

Deep engagement with these groups, whereby researchers in the groups were interviewed about their data needs and involved in User Acceptance Testing, was essential for developing a web portal that users can use with ease and which is capable of providing the data that they require. Although the software system cannot deliver all the RCM data that any given impacts researcher may require, the four collaborating research groups are able to use it to access the RCM data that is most fundamental to their research. This data is provided by the web portal in data files that can easily be read and manipulated by the collaborating researchers and should be sufficient to form the basis of research into the use of RCM output in the assessment of the impacts of climate change. The system is also likely to be able to satisfy many of the most fundamental RCM data needs of impacts researchers who were not involved in the project.

The software written as part of the CliMDDIR project has been designed to allow the CliMDDIR system to be easily extended. For example, although the software system currently serves data from a single regional climate modelling project, it was developed with future application to multiple different multi-simulation RCM datasets in mind. This makes it potentially valuable to anyone seeking to make RCM data more accessible and it is has been made available for this purpose through an open source repository. Elements of the software are also likely to be of use to anyone seeking to manipulate RCM (or GCM) output, such as researchers with a focus on the climate system itself.


Research Champion:

Prof Andy Pitman
ARC Centre of Excellence for Climate System Science

Data Type:

Input Data:
Output from Regional Climate Model simulations in netCDF format

Output Data:
1) Subsets of Regional Climate Model output in formats that climate change impacts researchers can use (e.g. ASCII, csv, GIS) 2) Data sets derived from Regional Climate Model output that are more relevant to specific climate change impacts than the raw model output (e.g .output averaged up to the temporal scales relevant to an impact, output with some statistical corrections or analysis applied)

High Level Software Functionality:

The CliMDDIR service allows approved users to access output from Regional Climate Model (RCM) simulations performed as part of the New South Wales and Australian Capital Territory Regional Climate Model (NARCliM) project. The service reads binary files produced by the project's Regional Climate Model (RCM) datasets that describe plausible future climate conditions at a local or regional scale have the potential to be hugely valuable to researchers investigating the possible impacts of climate change. However, RCM datasets are generally stored in a binary format called netCDF that is convenient for climate modellers to use but with which many impacts researchers are unfamiliar. The CliMDDIR service provides impacts reseachers with a user-friendly web portal that allows them to access impacts-relevant RCM data without the need for them to interact directly with specialist netCDF-handling software. The service has four key capabilities that facilitate the use of RCM datasets by impacts researchers:

Data subsetting
RCM datasets are typically vast, being derived from multiple multi-decade RCM simulations of regions that can be as large as a continent. The datasets comprise values of numerous climate variables saved at high temporal and spatial resolution for numerous different layers of the atmosphere. This data is invaluable for research focused on climate system processes and RCM evaluation but most of it is not directly useful to most climate change impacts researchers. Indeed, many impacts researchers are interested in daily mean average or daily total values of a few climate variables at or near the Earth’s surface for a either a selection of point locations or a limited area. These researchers face the challenge of extracting the data of relevance to them from an RCM dataset. This data subsetting is difficult without specialist knowledge of software designed to manipulate netCDF files. A key capability of the CliMDDIR service is therefore that it allows users to extract RCM data for the simulations, locations or regions, time periods and climate variables that are of interest to them via the web interface.

Data reprojection/interpolation
Climate change impacts researchers often require values of climate variables for a selection of point locations or for a grid that is either regular in terms of latitude-longitude coordinates or in terms of distance along the Earth’s surface (i.e. an equal area grid). RCMs usually output values on an irregular grid. The CliMDDIR service is capable of reprojecting/interpolating values on an irregular RCM grid to both point locations selected by the user and a regular latitude-longitude grid.

Data reformatting
The netCDF format in which RCM datasets are typically stored is generally not suitable for direct ingestion into the numerical and statistical models, or the analysis software, that many climate change impacts researchers use. The CliMDDIR service is capable of providing output in formats that impacts researchers are used to handling. Users can access gridded data as CSV (Comma Separated Variable) files or ASCII Grid files. CSV files are readable by most text editors and spreadsheet packages (e.g. Microsoft Excel). ASCII Grid files can be read by GIS packages. Site-specific data are available in a CSV format similar to that used by the Australian Bureau of Meteorology to distribute observational datasets that many Australian impacts researchers are familiar with.

Data exposure and re-use
The CliMDDIR service enables users to assemble data collections that can be exposed to the research community through the ANDS Research Data Australia website. A metadata feed containing descriptions of collections and their creators is provided by the service to the Research Data Australia website and this feed is used to generate Research Data Australia “collection” and “party” description web pages.

The CliMDDIR service has been designed to be easily extended. For example, it has been developed with future application to multiple different RCM datasets in mind. This makes the CliMDDIR software potentially valuable to anyone seeking to make RCM data more accessible and it is has been made available for this purpose through an open source repository. Elements of the software are also likely to be of use to anyone seeking to manipulate RCM (or Global Climate Model) output, such as researchers with a focus on the climate system itself.

ANZSRC-FOR code:

0401 ATMOSPHERIC SCIENCES