About the NSW DPI Combined Drought Indicator

For the latest news and information about drought in NSW, go to DroughtHub.

In 2008, a national review of drought policy recommended drought assistance be restructured to better help farmers plan and prepare for drought. Under the consequent Intergovernmental Agreement on National Drought Program Reform, the NSW Government established the Regional Assistance Advisory Committee (RACC) in 2013 to oversee the production of State Seasonal Updates to help landholders prepare for worsening seasonal conditions and advise the government on impacts of adverse climatic conditions.

The RAAC has worked to develop a consistent method of determining when farmers are being seriously adversely affected by drought, initially using a 'trigger' system based on remote measurements of rainfall, soil moisture and pasture growth, supplemented by on-ground data and observations provided by NSW Local Land Services.

NSW DPI’s 2015 review of seasonal conditions reports identified a need for a staged approach to seasonal conditions assessment using a comprehensive drought index (or indices) to describe both the onset of deteriorating conditions and recovery. Such an approach could deliver a graduated advisory system similar to the familiar fire danger warnings seen beside many roads in NSW. Each level of warning could be accompanied by a set of advisory actions and could support specific socioeconomic interventions.

The outcome of the review was the Enhanced Drought Information System (EDIS) project to track all phases of drought onset and production recovery. EDIS aims to:

  • build drought risk awareness
  • emphasise drought preparedness
  • improve confidence in drought monitoring and early warning, and move away from ‘in-crisis’ drought management.
While EDIS is focused on the co-ordination of Government services and support, it also has broad relevance to individual farm managers.

A key feature of EDIS is the development of the NSW DPI Combined Drought Indicator (CDI). The CDI integrates a range of complex data and model outputs in a framework that is useful for decision makers.

It combines meteorological, hydrological and agronomic definitions of drought using indexes for:

From these, a fourth index, drought direction (DDI), is developed.

Used together, these indexes can indicate the five phases of drought:

  • Non Drought
  • Recovering
  • Drought Affected
  • Drought
  • Intense Drought
Information from the CDI is used in NSW DPI’s Seasonal Conditions Report which contains official confirmation of drought conditions.

EDIS and the CDI have been developed in consultation with a range of stakeholders including:

Policy
  • NSW DPI
Technical review
  • Bureau of Meteorology
  • CSIRO Oceans and Atmosphere
  • International Centre for Applied Climate Sciences
  • Australian Bureau of Agricultural and Resource Economics and Sciences (ABARES)
On ground use
  • NSW Local Land Services
  • NSW Rural Financial Counsellors
  • NSW DPI Rural Resilience Network

Technical Information

The technical development and testing of the Enhanced Drought Information System and NSW DPI Combined Drought Indicator are described in a technical paper.

Used together, the following four indicies comprise the five phases of drought.

NSW DPI Combined Drought Indicator (CDI)

The NSW DPI Combined Drought Indicator (CDI) includes four indicators for rainfall, soil moisture, plant growth and drought direction which, used together, can indicate the five phases of drought: warning, drought, early recovery, strong recovery and non drought.

Rainfall Index (RI)

The RI is the percentile rank of rainfall aggregated over 12 months. The ranking is made using a 30 year (1980-2010) baseline which captures recent big shifts in climate variability, and factors in climate change. This provides an index between 0 and 100 where values approaching 0 are close to driest, and those approaching 100 close to the wettest, for any given region. Percentile-based indices like the RI have a uniform distribution regardless of their climatic setting, which is an attractive feature in NSW given the presence of rangeland, temperate and sub-tropical climates which have skewed, normal and log-normal rainfall distributions.

Maps showing historical rainfall index values in NSW

Soil Water Index (SWI)

The SI is calculated using the same procedure as the RI, but uses a soil moisture field derived from the DPI AgriMod soil water balance. Plant available soil water from layer one (0-10cm) and layer two (11-45 cm), the assumed maximum rooting zone, are aggregated and used to calculate the SWI. Similar to the RI, the SWI is an index between 0 and 100. In most districts of NSW a value of 0 means there is no plant available water held in the profile. The SWI is a hydrological index, but its configuration means that it is more useful as an indicator of conditions for dryland than irrigated agriculture.

Maps showing historical soil water index values in NSW

Plant Growth Index (PGI)

The PGI is calculated using the same general procedure as the RI, using the output from DPI’s crop and pasture models. Crop stress and pasture growth data are taken from DPI AgriMod, and the percentile rank calculated for each day. If the predominant land use in a given area is cropping, the PGI uses the crop-derived data, otherwise it uses the pasture growth indicator. The PGI is an agronomic drought index which is not only sensitive to moisture but also temperature variation and seasonal events such as frost. It is important to note that the PGI tracks the influence of climate on production potential across broad areas only. This provides a regional indicator of conditions. In the paddock, management decisions like fertiliser application and timing, sowing times and stocking rates drive outcomes on the ground, and in-field conditions can be above or below the regional indicator reported by EDIS.

Maps showing historical plant growth index values in NSW

Drought Direction Index (DDI)

The DDI is a categorical index that tracks the trend in meteorological conditions over the last 200 days. If the DDI signals a drying trend (DDI negative), and the other indicators are below the 50th percentile, a drought warning is issued. If the DDI signals a wetting trend and other indicators lift above the 10th percentile the region is in a drought recovery phase. The DDI is not a predictive index. Its values have a low to moderate correlation with what is happening in the immediate future. The DDI is derived by recursively fitting a robust linear regression function to the six month accumulated rainfall value each day, for a span of 200 days, and retaining the slope value of the regression. A robust regression is not overly sensitive to very high or low values or outliers in the data. The slope value is then rescaled to a range between -100 and +100 to normalise it for regional variation and produce the DDI.

Maps showing historical drought direction index values in NSW

Data Flow and Monitoring System

At a technical level EDIS is a data analytics and deployment system that sources and integrates climate and remote sensing data from a range of sources (Figure 1). Integration is achieved by an agricultural modelling data assimilation scheme (DPI AgriMod) and a series of algorithms that calculate individual drought indicators and the NSW DPI Combined Drought Indicator (CDI). New data for the CDI arrives every day, but this is for two days prior to the current date to allow time for data transfer from the field and quality control.


Technical diagram showing the workflow of how the CDI is calculated

Figure 1. Technical schematic of the Enhanced Drought Information System (EDIS) used to calculate the NSW DPI Combined Drought Indicator.

The algorithms used by EDIS have been custom-engineered to efficiently undertake computationally expensive tasks, such as running the data assimilation model or ranking current events in large data sets. Custom software engineering was required to manage the high computational burden involved in operating EDIS as well as to complete the science behind its development. The baseline climate data sets have daily temporal resolution, whereas the remote sensing data has high spatial resolution (down to 250m²). Operational runs of EDIS need to be completed each day in a reliable and timely manner, within a 30-60 minute processing time. The experimental tests of EDIS described in the technical paper involved repeated daily calculation of the CDI given different configurations of the framework across the 50-100 year climate record.

Frequently Asked Questions

How is the CDI linked to seasonal conditions?

The CDI weekly summaries of drought conditions in Local Land Service regions are compiled each Thursday by the NSW DPI Seasonal Conditions coordinator using CDI regional maps. At the end of each month the CDI information is verified and quality controlled for use in the State Seasonal Updates to provide the official drought status summary for the state.

How is the CDI linked to drought support?

Through the Seasonal Conditions Report information from EDIS is considered by the NSW Rural Advisory Assistance Committee (RAAC). Refer to the DroughtHub for more information about support available to NSW farmers.

Does the CDI provide information about a specific farm?

CDI currently provides information down to a parish level, which is a composite of 10-20 farms. The underlying monitoring of conditions occurs on a 5km grid, which is a composite of 3-5 farms. Work is underway to monitor conditions on a 1km grid, around a single farm. Information is not reported at an individual farm level to maintain commercial confidentiality. It is available upon request, and work is underway to develop a Farm Drought Forecast service.

Why is drought monitored on a 5km grid?

Currently the best available data for tracking climatic condition across NSW are from the Australian Water Availability Program, which was developed by CSIRO and now run operationally by the Bureau of Meteorology. The limit of accuracy in this data is 5km, which is about 3-5 farms. The Australian National University has developed techniques to track climate on a higher resolution 1km grid, which is about 1 farm. DPI is working with Australian National University and University of Sydney through the Terrestrial Environmental Research Network to bring this into operational use.

How does the CDI classify drought status within a parish?

A conservative approach is built into the classification of a parish’s drought status. If any part of the parish is experiencing drought conditions, the whole of the parish is deemed to be under drought status. This also applies to issuing a warning. All of the parish must be experiencing drought recovery or non drought conditions for the area to be in these phases.

Can the CDI predict or forecast drought?

Currently the CDI does not predict or forecast drought. While the Drought Direction Index (DDI) indicates a wetting or drying trend over the past 200 days it is not a predictive index. It is best used for categorising a recovery or warning phase. Testing shows that when in a warning, more often than not rainfall occurs that lifts conditions and a region does not necessarily descend into drought. Work is underway to integrate a numerical weather forecast modelling system with EDIS.

Obtaining EDIS Data

EDIS data can be obtained by establishing a third party license agreement with the Department of Primary Industries. Depending on data use, administration and commercial fees may apply.

For more information contact edis.info@dpi.nsw.gov.au.

Acknowledgements and Disclaimer

The EDIS portal is built, maintained and hosted by Intersect

The data supporting EDIS is made available by third parties. Special acknowledgement is given to the Australian Bureau of Meteorology for provision of the Australian Water Availability Program (AWAP) and AVHRR NDVI data sets.

Other third party data sets are made available through the Terrestrial Environmental Research Network (TERN) under the Creative Commons – Attribution 4.0 International (CC BY 4.0) licence. Attribution and ownership of each data set are listed in the following table.

ClassItemSourceProperties
Climate Australian Water Availability Program (Rainfall, Max/Min Temperature, Radiation, Potential Evaporation) BoM Gridded, 5km² resolution, daily
ANU Climate TERN/ ANU Gridded, 1km² resolution, daily
Remote Sensing MODIS and derived products
Fraction of photosynthetically absorbed
Radiation (fPAR) (1km)
Leaf area index (LAI) (1km)
National Dynamic Land Cover Dataset (250m)
Actual Evaporation (250m)
Vegetation Indices (250m, 500m)
Gross Primary Production (250m)
CSIRO/ TERN/ GeoScience Australia Gridded, various resolutions, dekads and months
AHVRR
Normalised Difference Vegetation Index (NDVI) (1km)
BoM Gridded 1km² Dekads
Soils Australian Soil Landscape Grid CSIRO/ USyd/ TERN Gridded, 90 meter

EDIS and CDI are trademarks owned by the NSW Department of Primary Industries, an office of the State of New South Wales through the Department of Industry, Skills and Regional Development. Copyright in EDIS and CDI vests in the State of New South Wales through the Department of Industry, Skills and Regional Development, 2016.

Any use of EDIS, CDI or any related material must be under written licence on terms approved by the NSW Department of Primary Industries. Administration and commercial fees may apply.For more information contact edis.info@dpi.nsw.gov.au

Disclaimer: The information contained in this web portal is based on knowledge and understanding at the time of development. Because of advances in knowledge, users are reminded of the need to ensure that information upon which they rely is up to date and to check currency of the information with the appropriate officer of the Department of Primary Industries or the user’s independent adviser.

Contacts

EDIS data

If you are interested in obtaining EDIS data or initiating a development project, contact:
edis.info@dpi.nsw.gov.au

Seasonal conditions reports

EDIS communications officer:
seasonal.conditions@dpi.nsw.gov.au