A network of weather stations in South Australia to detect thermal inversions and help prevent pesticide spray drift
Hi from the team who built the "mesonet" weather station networks in the state of South Australia!
A mesonet is a term for a network of automated weather stations designed to monitor meteorological phenomena at high geographic density and updated frequently.
We built these networks to help prevent pesticide spray drift - the phenomenon in which pesticide spray that is applied to a crop ends up drifting to other crops, sometimes many kilometres away.
Spray drift can be a huge problem, as it can lead to damage or even destruction of highly valuable crops, and can also cause pollution to waterways and harm to native ecosystems.
The biggest factor in spray drift the presence of a thermal inversion, a state in which temperature increases with increasing altitude, which usually happens in the evening and overnight when the earth is cooling, but can happen at other times. However other meteorological phenomena also affect the likelihood of spray drift, and different types of spray respond differently.
So we designed a weather station to measure all the relevant meteorological phenomena, and in particular to detect thermal inversions, and a custom web app to display the data in format that is very fast and simple to access and understand.
We've now deployed 70 of these stations across two of the major agricultural regions of South Australia, and so far it seems to be making a difference: no damage due to spray drift has been reported in these regions since the networks were rolled out.
Background
A group of stakeholders in the Mid North region of South Australia had identified spray drift as a major issue.
The region, to the north of South Australia's capital of Adelaide, has a thriving agricultural sector, producing wheat, barley, wine grapes, pulses, livestock and other produce.
The total annual output from the region is valued at nearly AUD $2 billion. It has been estimated that the potential loss in production from spray drift could be as high as $178M/year, and potential harm to waterways and native ecosystems is significant.
The local stakeholders worked with meteorologists and researchers to develop a solution, then sought funding from the South Australia state government to fund the project.
In 2018, a $1.4M grant was provided, and a pilot rollout of 40 weather stations was undertaken. It was named the Mid North Mesonet.
The pilot was deemed successful, and in 2019, another grant was provided by the state government for a network in the Riverland and Mallee region to the north-east of Adelaide.
In the two years since the Mid North Mesonet was rolled out, there have been no reported instances of crop damage due spray drift.
The Problem of Spray Drift
Farmers spray pesticides on their crops to manage pests. It is an important feature of modern farming. Herbicide may be applied to target broad-leaf plants, other times to target summer weeds.
Spray drift occurs when pesticide spray 'drifts' across into non-target areas.
Spray drift is a problem because:
- The affected non-target area may be particularly susceptible to pesticides (i.e. broad-leaf crops such as grapevines are especially vulnerable)
- Some markets are highly sensitive to the amount of pesticide residues detectable in the end products (i.e. wine exports to China; organic farms; etc)
- The affected non-target area might be a sensitive natural ecosystem (i.e. local waterways)
- The affected non-target area might include rain water tanks used for drinking water
- It is a waste of pesticide product and labour time costs
Causes of Spray Drift
A combination of conditions can lead to spray drift.
Wind Speed
Wind speed that is very low or very high.
Temperature
A temperature over about 28°C (83°F).
Humidity
The effect of humidity varies depending on the type of spray. The spray manufacturer's instructions should provide guidance.
Atmospheric Stability
Unstable or stable conditions can lead to spray drift. Conditions should be neutral.
Thermal Inversions
A thermal inversion, in which temperature increases with increasing altitude, is a major risk factor for spray drift.
Inversions generally happen in the evening and overnight as the earth cools, but they can happen at other times.
Thermal Inversions
For current farming practices, long-distance spray drift generally occurs during very stable weather conditions (i.e. little to no air turbulence).
A thermal inversion occurs when a warm layer of air sits above a cooler layer of air near the ground.
This reduces air turbulence and can act as a �~@~Xlid�~@~Y for an airborne pollutant source. The lack of air turbulence means the smaller spray droplets float without settling on the target crops.
A very slight breeze can then carry these floating droplets large distances before they eventually descend into a non-target area.
It is illegal to spray during a thermal inversion.
The problem: it is very difficult for farmers to tell if a thermal inversion is underway.
Our Weather Stations
Our weather stations were designed to:
- Measure temperature at 1.2m and 10m, in order to detect inversions
- Measure all weather metrics that are relevant in determining whether it is safe to spray
- Update data every 10 minutes, so farmers have access to the most recent readings
They consist of:
- Temperature sensors at 1.2m
- Temperature difference between 10m and 1.2m
- Wind speed and direction sensors at 2m and 10m
- Tipping bucket rain-gauge
- Solar radiation sensor
- Pressure sensor
- Relative Humidity sensor
- Antenna
- DataLogger which takes readings every 10 minutes, does some calculations locally, and uploads over cellular data (3G/4G) to the web server for processing and display
Each tower is self-powered by a solar panel and battery.
The Web Application
We initially used an off-the-shelf web application for displaying meteorological data, but soon realised we needed something tailored to our needs. We found a company that was already in the business of building web-applications to display environmental data for farmers, and they were willing to build a customised version for us.
The key requirements for the application were:
- Each of the monitored metrics displayed on a map, with fast/simple switching between each metric
- A dashboard of all the current key metrics and recent historical graphs
- Long-term graphs of all the key metrics also available, and as well as the past 48 hours of data in tabular format
- The site must display well and be easy to use on a mobile device, and be fast to use even in areas with low cellular signal strength, given that much of the usage will be by farmers out in the field in remote areas.
The key technologies are:
- Ruby on Rails
- PostgreSQL
- Redis
- Sidekiq for queue processing
- React
Data is uploaded from the data logger via FTP (which may seem primitive, but the world of meteorological data still relies heavily on FTP).
On detection of new data, the application first takes the most recent readings, and updates the data structures for the map and dashboard views.
The historical data is then placed in a queue, for updating of historical graphs.
All the data for the map view, dashboards and historical graphs are formatted into a JSON hash then serialized and stored in the Redis cache. That way, when the React web interface makes a request, the pre-formatted data can be retrieved and sent very quickly, as no querying, processing or formatting of data is required at request time.
Results so far
In the two years since the Mid North Mesonet was deployed, there have been no reported cases of crop damage due to spray drift.
Damage and loss had been reported in most of the previous years - though not all. So it is too early to tell if the problem has been completely solved, but indications are promising.
Other government bodies and NGOs are expressing interest in establishing their own mesonets.
Contacts
Damon Grace
Project Engineer
damon.grace@cotl.com.au
Warwick Grace
Meteorologist
warwick@graceresearch.com
Tom Howard
Web Developer
tom+mesonet@taindata.com