1 Agriculture Western Australia:
3 Baron-Hay Ct., South Perth, WA, 6151
Phone: (08) 9368 3272, Fax: (08) 9368 3939
2 Intergraph Corporation Pty. Ltd.:
U2, 4 Brodie Hall Drive, Bentley, WA, 6102
Phone: (08) 9472 3266 Fax: (08) 9472 3267
The latest in cutting edge computer technology played a key role in controlling the threat of Australian Plague Locusts (APLs). The technology was put to the test during the recent locust outbreak which was predicted to be at least twice as large, both in numbers and area, as the last incident in 1990 which caused $20 million damage. With surveys indicating the outbreak was going to affect two thirds of shires in the agricultural region, monitoring the incident and coordinating the control campaign was a mammoth task.
An integral part of the agency’s emergency response was the development of a computer database and mapping program to track almost every facet of the campaign. Working closely with Perth staff of computer mapping company Intergraph a specialized locust application was developed based on its Geomedia Web Map technology. The comprehensive Locust Web Map application provided details of locust hatchings, surveys and the spraying program at a Statewide, regional, district and property level.
The Locust Web Map tool has enabled the control teams to make more timely and informed decisions about if, when and where to undertake locust control.
It assisted in the coordination of more than 400 staff and 500 landholders across more than 550,000 hectares. It was the first time Agriculture WA (AGWEST) had used such technology, which could also be applied to other agricultural emergency situations, like an outbreak of an exotic pest or disease.
Background to the Outbreak
Why did the agency need to develop a GIS Application for this outbreak?
The reasons are mainly related to the size of the locust’s populations and to the widespread nature of the incident.
A survey of locust numbers was conducted in 56 Western Australian Wheatbelt shires in April and May 2000. Analysis of the data strongly indicated the likelihood of an extensive locust outbreak occurring in the spring and summer of 2000 over a large part of the Southwest Land Division. The data also indicated that the outbreak would be more severe than the 1990 outbreak, which was the worst experienced in WA.
Emerging nymphs and swarming locusts have the ability to cause severe damage to crops and pastures in the Wheatbelt as well as damage to intensive horticultural enterprises, sporting grounds, orchards, vineyards, gardens and other public and private facilities.
Previous outbreaks of APLs had occurred in the Southwest Land Division. The most extensive recent outbreak occurred in 1990 and the response cost the State Government approximately $4.2 million. Damage to agricultural and horticultural industries was estimated to have been approximately $20million.
- Secondly, despite early planning a limited amount of chemical was going to be available. Thus the target spraying zones need to be chosen in order to ensure that the maximum number of insects would be sprayed.
- And thirdly, the environmental requirements have changed in the last decade so that all spraying actions needed to be well-documented enabling prompt answers to any incidents reported by the public. In addition, the far more diverse nature of agricultural enterprises meant that it was important to define appropriate buffers around sensitive areas and to avoid organic farming enterprises.
Planning of the control operations commenced in July. This planning involved issues associated with communications, resource requirements (human, physical, and financial), staff briefing, contractor engagement, chemical availability and distribution. Also there were mapping, database and geographic information system requirements which became the responsibility of the Spatial Resource Information Group (SRIG), a unit within AGWEST.
The field officers who were going to carry out inspections of properties and advise on spraying targets needed accurate hard copy maps showing recent farm infrastructure. The taskforce in general needed databases that collated when hatching occurred, farmer reported high insect counts and finally the insect counts carried out by AGWEST field staff as part of formal monitoring.
All the databases needed to be related to property boundaries and displayed in a geographic information system so decisions could be made on target zones for spraying.
Agriculture WA client property event system
The existence in AGWEST of an information system called Client Property Event System (CPE) has greatly enhanced AGWEST’s ability to handle emergency situations.
CPE (Beeston et al 1999) utilises the States spatial cadastral database parcel information, which is then integrated into properties using data gathering during the Agency’s daily activities. An agricultural property is thus defined as a series of contiguous cadastral parcels being managed as single enterprise.
The CPE database includes information about properties, owners and managers, addresses and information about the Agency's dealings with agricultural properties. Other systems currently linking to it relate to treatment of disease and protection of agricultural land. These include the livestock and disease information system, stock brands registry and field reporting systems.
CPE provides the technical infrastructure for efficient data storage and delivery of information in a variety of graphical and textual formats. In addition, CPE data can be spatially represented and analysed with many spatial data sets because it is integrated with the agency geographic information system. Delivery of all data and information products can be through stand-alone workstations or web based applications.
The updating of the property intelligence is a continuous task and essential to allow immediate and effective use in emergency situations. Property and client details for 48 shires were updated and entered into CPE database during this campaign.
Geographic Information System requirements
While the various requirements will be dealt with separately below it should be noted that all development proceeded in parallel. Also, the integration of the system components was vital to the success of the operations.
An analysis was made of data requirements, data collection requirements and data recording procedures and this identified that data collection and recording was to occur in two distinct areas-locust numbers and distribution and application of control measures.
Information about the locusts was collected in the field by farmers, volunteers and AGWEST officers and recorded on hard copy forms which were collated at the nearest AGWEST district office (DO) and entered onto the databases by the local Data Entry Officer (DEO’s).
There were four official sheets;
District Locust Hatching Summary – data received from farmers or AGWEST staff indicating hatchings have been sighted on a property.
Local Area Summary Sheet (LASS)– estimated counts of locusts on a property by farmers or volunteers.
AGWEST Verification sheet (AVS) – estimated counts of locusts on a property by trained AGWEST staff either as a verification of the LASS or as a first survey.
Property Survey Sheet – revised density of locusts on a property by paddock and prescribed spraying.
The information from these sheets had to be entered into a database in order to be mapped. Two databases were created for this purpose. The Hatching, LASS and AVS were recorded concurrently during the campaign and so were created as one database. The Property Survey Sheet was not used until later in the campaign and was sent out as a separate database some time after the first.
Both databases were made as easy to use as possible. The data entry forms were made to look as much like the corresponding hard copy form by embedding the word documents as objects. They were also produced as stand alone databases for each DO which were emailed to head office daily. This avoided any potential network problems which might have been caused by a using a central database as well as automatically creating a daily backup of each DEO’s database.
During the height of the campaign data was received from DEO’s from 11 different district offices twice daily, 7 days a week. Over 1000 database files were received from country centers and processed into a central database.
Updates were made to the Web Map application at least once a day and during the major weeks of the incident on two occasions per day. All locust reports required a property number in order to be correctly mapped on the Web Map. In most cases, property numbers (unique ID’s in the CPE system) were known for a property. Thus, the information on the databases could be loaded using a series of Visual Basic scripts, incorporating ActiveX Data Objects (ADO) into the Master database.
In a relatively small number of cases the property number was not known or was incorrect. Unmatched properties were extracted to another database during the loading process by the Visual Basic scripts using ADO and then located and corrected manually by SRIG staff.
Each of the forms (Hatching, LASS, AVS and Property survey) were used to create one or more themes on the web map.
Hatching map – shows distribution of locust hatching reports.
Lass & avs
Surveyed map – surveyed maps shows colour coded properties depending on survey type i.e. Farmer (lass), lass & agwest staff (avs) or only avs.
Lass & avs
Risk map – colour coded properties indicating the number of locusts in millions calculated from density estimates in the database and the property area in cpe.
Prescribed sprayed areas.
Web Map Application
The Master database described in the previous section was incorporated into a Geomedia Warehouse once it had been matched to the properties in the CPE System. This Geomedia Warehouse then was web enabled using Intergraph Geomedia Web Map into the Locust Web Map application. This application was developed by AGWEST staff working closely with Perth staff of computer mapping company Intergraph.
GeoMedia Web Map is a web based product that publishes intelligent maps from GIS databases for Intranet or World Wide Web use. This product operates under Microsoft’s Internet Information Server (IIS) and uses Active Server Pages (ASP). This environment provided easy and open access to AGWEST Locust and Mapping databases through the use of both GeoMedia Web Map COM Objects and Microsoft ActiveX Data Objects (ADO).
The delivery of mapping and attribute information over AGWEST’s Intranet to the regions was via Internet Explorer using an ActiveCGM plugin. The easy access to AGWEST data and information delivery through the AGWEST Intranet was one of the critical elements in ensuring that AGWEST regional offices received updated information in a timely manner. GeoMedia Web Map also provided the powerful spatial query tools in the Locust Web Map application, such as being able to draw area of interest on the map in Internet Explorer and find locust counts on properties for the given area.
The early predictions of the outbreak extent indicated that a farm plan might be needed for up to 9000 properties in the Agricultural Zone. Not all properties in the area had carried out farm planning in the Landcare process. Also some of the farm plans were out of date and did not show current infrastructure.
A decision was made to utilize a new series of black and white orthophotographs prepared from 1:40,000 photographs flown in the 1997-2000 time frame. These orthophotos were combined with property boundaries derived from the CPE system utilizing Microstation and I/RAS C software. Depending on the size of the property they were printed at scales between 1:10,000 and 1:25,000 with the property centroid also shown.
These field maps were used by AGWEST staff when visiting properties to identify all sensitive areas (houses, dams, grain storage and aquaculture sites) and to identify the target areas for aerial or ground spraying. Farmers also gave permission on these documents for spraying to be undertaken.
The aerial spraying operators took these maps into the air with them and used them for navigation and the property centroid to set their Global Position in moving from one farm to another. Ground operators also utilized the maps for navigation and at the completion of spraying all the target areas were captured utilizing Microstation.
The size of the task that confronted staff during the incident can be gauged by the volume of mapping outputs. These included a total of 4020 property maps and 125 mapsheets at a scale of 1:50,000 composed of 2012 orthophotos. In addition where orthophotos were not available a total of 1235 colour aerial photographs were rectified to fit topographic mapping and all prescribed sprayed areas have been entered into the original district maps for matching with spray tracklogs.
Processing of tracklogs
A requirement of the contract let by AGWEST to aerial spraying contractors was that all spraying must be recorded by a track log system. These files were then processed by propriety software and displayed over the digitized approved spraying areas. Tracklogs for all aerial spraying are in process of being prepared into maps.
In an organisation as large and diverse as AGWEST there is considerable variation in the levels of general computing experience and confidence. This became more obvious during the locust campaign and highlighted the need for clear, simple instructions as well as easy to use databases and applications. Procedure manuals were prepared for all aspects of the mapping and database operations and these are being collated for use in any future operations of the same type. In some cases manuals were rewritten several times to make them easier to understand.
It also became clear early in the campaign that written instructions were not always enough to get taskforce members with little computing confidence up and running, so support was essential. This was achieved by nominating a SRIG staff member at each regional office as the supporting officer. A list of these officers and their contact details was added to every procedure manual and mail out.
In many cases taskforce members were found to be using software and hardware that was considerably older than the recommended agency standard. Installation of later versions of Internet Explorer and Microsoft Access was required and some officers required new monitors in order to have the necessary screen resolution to view the Web Map. Ensuring machines or software are upgraded to a standard on a more regular basis would improve the speed of implementation for any similar campaigns in the future.
The program eliminated the need for pins on the wall to monitor the progress of the outbreak and the control campaign. As the program was posted on the agency’s internal web site, it was able to update staff about the statewide situation. When compared to the previous locust campaign this was shown to have considerably reduced time and phone calls between regional offices as they compared maps. It also reduced tensions between campaign regions as all regional officers could see the most severely effected areas and the reasons for chemical deployment to regions other than their own.
While new technology played an integral part of the agency’s emergency response it would not have been possible without the dedication of the AGWEST staff involved in all aspects of the incident. While only three members of the staff are shown as authors of this article every member of SRIG contributed to the utilization and success of the technology described.
Beeston, G.R., Kabay, M., and Steward, K. (1999) Towards an integrated data model for agricultural enterprises in Western Australia. WALIS forum ’99, Perth W.A., pp 44 - 47.