Data Services

Mapping Wildfires in Southern Belize

Mapping Wildfires in Southern Belize – the case for analysis ready data (ARD)

Savanna based ecosystems across the world are suffering increasingly from wildfires due to climate change and illegal human activities. In many regions, this puts the fragile ecosystems under threat, so mapping the extent of wildfires becomes important to enable the organisation of mitigation measures.

Recently, Environment Systems supported a Masters student from the University of Edinburgh School of GeoSciences, Chris Halliday, in a project that sought to investigate a new approach to mapping savannas. An 1,800 km2 area of Southern Belize was chosen as the area of study. The area was chosen because it suffers from extensive wildfires, which destroy saplings, the habitats of nesting birds, and cause a general decline in biodiversity. Three ‘Protected Areas’ within this area are designated to protect key savanna species.

Currently, the burnt areas of savanna are mapped annually at the end of the dry season in May by visual interpretation of Sentinel-2 optical imagery. This method requires cloud-free imagery, which is not always available. In addition, the timing is not optimal due to rapid savanna regrowth. Radar data, which can penetrate cloud, is not generally used to map burnt areas of savanna as few land managers have the required expertise to handle this data source.

Burn areas in Southern Belize
Burnt areas mapped from Sentinel-2 (left) compared to burn areas mapped using a time-series of Sentinel-1 indices
Step-in Sentinel-1 analysis ready data (ARD) from Environment Systems Data Services. The project investigated pairs of radar images before and after a fire. The physical basis for detecting burnt areas using radar relies on being able to observe changes in backscatter over time. With imagery captured from January to December 2019, object-based image analysis was used to compare radar- based methods with the visual analysis of Sentinel-2 imagery obtained for the nearest dates. The radar-based method detected 87.6 % of the burnt areas compared to the visual analysis, but was also able to reveal more about fire evolution over the season due to the increased frequency of the data capture, and its ability to see through the cloud.

Monitoring Grassland By Satellite

Back in 2019 we were involved in a pilot project which investigated the potential of using satellite remote sensing to inform grassland management and predict grass yield in Wales. Things have moved on since then but the reasons for wanting to do this have not really changed. Grass is an essential crop in livestock production, and grazed grass is the cheapest and most efficient form of feed on a farm. When managed well, inputs and production costs can be reduced, boosting profit margins.

Measuring and monitoring grass growth enables the farmer to improve quality and maximise yield, and make decisions about stocking, grazing rotation and fertiliser applications. On smaller farms this is achieved with a rising plate meter which measures the growth and quantity of available ‘Dry Matter’ per hectare (DM/ha). It is a very labour intensive and time-consuming process so the attraction of using data from satellites which pass over the farms every 6 days is plain to see. In the pilot project Environment Systems developed an algorithm that can predict average grass cover (Kg DM/Ha) to develop a ‘hands free’ online tool to help farmers estimate average grass cover. The algorithm uses radar data, chosen for its ability to penetrate cloud cover and then provide consistent data readouts over time.

Plate meter versus satellite data
Fields measured and modelled track well over time
Since that time, we have proved that the technology is transferable to other regions. We have been using radar data in Colombia where persistent cloud cover prevents the use of more traditional optical satellites. The consistent source of data over time is helping farmers to manage their grazing grassland more efficiently.

We have also proved that the technology is scalable. We have been successfully monitoring grassland on a 50,000ha farm in New South Wales, Australia. These are exciting developments because we now know that our technology is both transferable and scalable which means that it can be applied to farms previously considered too large for routine use of a plate meter. Starting out in a small pilot project in Wales less than eighteen months ago we have now proved the efficacy of satellite data for monitoring grassland within a single field or across entire continents!

In November Caron Pugh one of Senior Consultants presented the findings of our research in the Precision Livestock Farming And Sensing Technology In Extensive Grassland Systems Webinar run by the BSAS (British Society of Animal Science). His presentation is now available on YouTube or you can access the whole webinar here.

One Million Fields Surveyed!

Incredible as it may seem, during one of the weekly updates between members of our Data Services Team they calculated that we have surveyed over one million fields in the last year. When we take into account the number of repeat visits to some of these fields, the land area we have covered is well over 600 million km². We’re always busy in the UK but more recently we have been expanding our focus to include Latin America. We have covered huge areas of soybean cultivation in Brazil and Argentina, 70,000 km² in Peru and 68,000 km² in Colombia for the EO4cultivar project. In addition, under the project extension, we have also surveyed areas of Paraguay, Ecuador, Venezuela, Honduras, Belize, Dominican Republic, Jamaica, Cuba and Mexico.

In most cases we are mapping commercial farms. We have become quite the experts in banana plantation maps for example! When we say mapped it is more accurate to say classified and monitored. In many instances no maps exist and we have become very adept, using our own algorithms, in creating field maps so that farmers and their supply chains can readily identify and deal with issues as they arise. Multi-temporal monitoring of crops and identification of growth stages is beginning to revolutionise the way that growers manage their assets.

oat classification
Oat crop classification, Saskatchewan, Canada
This year we have also moved at scale into North America. Our work has specifically focused on oat growing regions in the US and Canada. We have mapped over 580k parcels of land in North Dakota, New Brunswick, Quebec, Ontario, Alberta, Saskatchewan and Manitoba providing crop growth stage intelligence in easily accessible online data dashboards.

Sentinel-1 and Sentinel-2 satellites provide the majority of the data we process but we are also avid consumers of Planet and other commercial sources, with much of the analyses automated using our own algorithms developed in-house.

Three crops we’re mapping from space!

Feeding the World…sustainably

With the world’s population projected to grow to approximately 10.9 billion by 2100 and with the recent COVID-19 pandemic, the world’s food system is under increasing pressure. We need to plan to feed more people without overwhelming the planet.

Satellite imagery can help safeguard global food supply and support sustainability. It provides a means to:

  • Assess fields at both local and national scales
  • Monitor growth uniformity
  • Identify weeds, pests and diseases
  • Diagnose nutrient and/or water deficiencies
  • Delineate and quantify weather damage (e.g. wind damage)

Environment Systems has been working with the agricultural sector for many years, providing data to governments, research institutes and growers. Here we touch on just three of the many different crops we are mapping from space:

  1. Oil Seed Rape (OSR)
  2. Bananas
  3. Soybean

1. Oil Seed Rape

The familiar, bright yellow oil seed rape crop is grown for the production of animal feed, vegetable oil and biodiesel. In the 1970s, the UK produced a few thousand tonnes. Today, 2.2 million tonnes of OSR are produced annually in the UK!

Despite being an important break crop, OSR is not particularly sustainable. This is because it is extremely vulnerable to a large number of pests and diseases which are generally treated with a range of fungicides and insecticides. Furthermore, in order to maximise yields and profitability, farmers will often use high applications of nitrogen fertiliser (increases risk of nutrient pollution in local waterways). However, knowing where OSR is can help land managers to reduce risks of water pollution.

Environment Systems has developed a method for identifying oil seed rape fields that uses a time-series of Sentinel-1 and Sentinel-2 satellite imagery. This makes it possible to estimate yields/prices, identify areas (waterways) at risk from OSR production and track production at local and national scales. You can view some of the areas we have mapped in 2019 and 2020 by viewing the map below.

Our OSR products for 2019 and 2020 are available to purchase at £0.40/km². Find out more here, or contact us to get the data!

2. Bananas

Banana Plantation
Bananas are the world’s favorite fruit in terms of consumption quantity, and are one of the most important staple foods. In addition, increasing health consciousness has led to increasing demand for bananas.

Banana plants are perennial (live more than two years) and take 9 to 12 months to produce fruit ready for harvest. For this reason, banana plants are not set in a rotation but are commonly grown in large plantations across Latin America. This requires large investment in infrastructure and technology to support irrigation, harvesting and transportation of the bananas.

Banana plantations will usually have high levels of in-field variation; a result of the different ages of the plants, environmental factors and management practices. A time series of satellite imagery supports banana managers to identify low performing areas and target management.

Environment Systems is currently mapping bananas in 14 countries covering an area of 9 million hectares! We are also providing crop health analysis to banana growers to support their management. Read more here.

3. Soybean

Soy Bean Plantation
Soybean is a legume, native to East Asia, that is a high-protein food consumed by both humans and animals. It has been grown for centuries although over the last few decades, production has increased dramatically due to its use in animal feed and an increased worldwide demand for meat. More recently, soybeans popularity as a ‘superfood’ has further contributed towards its demand.

The growing demand for soybean has led to an increase in deforestation in LATAM. For example, the Chaco Forest which extends over parts of Argentina, Paraguay and Bolivia is considered a biodiversity hotspot. Earth observation provides a means to map and monitor such activities.

Environment Systems has developed a method for mapping soybean fields that uses a time-series of Sentinel-1 satellite imagery. With an accuracy of 95%, we have mapped over 197000 km² in Brazil and Argentina – that’s almost ten times the size of Wales!
Read more here.

Forests 2020 – Soy Mapping

Forests 2020 is a major investment by the UK Space Agency, as part of the International Partnerships Programme (IPP). Project managed by Ecometrica, Forests 2020 uses advanced mapping technologies, satellite data and other insights to help protect and restore tropical forests through improved forest monitoring. Ecometrica has partnered with Environment Systems to deliver an automated and scalable approach to mapping soybeans in Brazil. Brazil possesses about one third of the world’s remaining rainforests, covering almost 60% of the country’s surface. It is also the largest soybean producer and expected to export 77 million tons of soybeans in 2020.

Western Bahia was identified as the pilot area, a region of Brazil characterised by large scale intensive agriculture, the majority of which is soybean cultivation. Here the rainy season stretches from October through to April, during which time soybeans are the principal crop, being sown from October and fully harvested by the end of March.

Soy Classification
A section of the classification in the pilot area showing areas of soy and non-soy
On this project we used our own Data Services to acquire the necessary Sentinel satellite data. The field level mapping approach required field boundary data. This was created using our own machine learning algorithm which automatically generated over 18,500 field boundaries covering an area of over 80,000 km². Radar satellite imagery was also acquired for the 2019-20 growing season generating a ‘soybean profile’ for each field showing how it grows over time.

To develop and test the map and achieve higher levels of crop classification accuracy, field work was required to verify the crop on the ground and thus provide additional ‘training’ points for the algorithm. Environment Systems surveyors collected over 1,100 data points. These points were applied to the field boundaries and satellite data to produce an automated classification with an accuracy of over 90%.

Following on from the pilot, the approach was rolled out in another of Brazil’s intensive agriculture regions, Mato Grosso. Almost 45,000 field boundaries were generated and an area of 100,000 km² was successfully mapped, identifying all the soya production, at a field scale, without the need for any field surveys. The maps are now accessible online via Ecometrica’s EOLab platform. You can find out more about Forests 2020 here.

Space-based Solutions in Agriculture Prove Their Worth

A recent report prepared by London Economics and Caribou Space, for the UK Space Agency (UKSA), has found that space-based satellite Earth observation solutions for agriculture can be six times more cost-effective than non-space alternatives such as drones, field patrols, and extension workers tasked with providing farmers with training and support.

The UKSA’s International Partnership Programme (IPP) is a 5 year, £30m per year programme, focusing on using the UK space sector’s research and innovation strengths to deliver sustainable economic and societal benefits, to emerging and developing economies around the world. The programme includes a number of projects focused on agriculture. Space- based solutions are particularly suited to addressing some modern challenges such as increased population, climate change, and poor yields in staple crops such as wheat and rice. Space-based solutions cover three specific areas:

  • Decision support tools
  • Early warning systems-early detection and mitigation of events such as drought
  • Farming credits, based on data collected for farmers in the developing world
ICA Region Peru
Near infrared satellite imagery – ICA Region, Peru

The cost-effectiveness in the report is measured in terms of the absolute value of the change in crop yield (£), and is dependent on the relative costs of data collection. The forward-looking analysis shows that agri-businesses can act on the intelligence derived from satellite data on yield or risk, whereas non-space solutions rely on farmers responding to guidance.

Environment Systems is proud to have been leading EO4cultivar, one of the 33 projects run in 44 countries since the IPP’s inception in 2016. The project has recently been expanded, see ‘EO4cultivar Expansion’ below. The delivery of satellite-derived agricultural data to agri-businesses, supply chain organisations, and smallholders is addressing risk management, supply chain visibility, crop yields, crop input management and showing great promis

Mapping Grassland by Satellite

This pilot project is establishing the potential of using satellite remote sensing to inform grassland management and predict grass yield in Welsh pasture systems. Why might you want to do this? Grass is an essential crop in livestock production. Unfortunately, around half of the grass grown in Wales is not utilised efficiently. Grazed grass is the cheapest and most efficient form of feed on farms and when managed well, inputs and production costs can be reduced, therefore boosting profit margins.

Measuring and monitoring grass growth enables farmers to improve quality and maximise yield. As part of the ongoing Welsh Pasture Project, a number of farms are using a plate meter to measure growth and quantity of available Dry Matter per hectare (DM/ha) across their farm, to monitor growth during the season. There are varying degrees of enthusiasm for this activity as it is labour intensive and very time consuming. There is an opportunity to use satellite data, with Environment Systems developing algorithms that can predict average grass cover (Kg DMA/Ha) with the eventual aim of developing a ‘hands free’ online tool to help farmers estimate average grass cover and reduce the need to go out into the field and collect data.

Grassland Mapping

The project utilised farmers plate meter data collected across nine cross-sector farms, located throughout Wales. To date, we have achieved positive correlations with the satellite radar data. The image (above) shows field parcels with grass cover modelled from satellite imagery, for 15th September 2019. The chart on the left shows how well our grass cover model (orange) matches actual data (blue) over time.
“The technology has the potential to support farmers across Wales to make informed decisions on grass management based on actual grass growth and data. Additionally, this information will be able to provide trends on individual field performance over a number of years therefore allowing farmers to compare against previous years and other fields on the farm. This information could help in deciding which fields to reseed based on performance, therefore avoiding unnecessary reseeding expense,” said Dewi Hughes, Technical Development Manager for Farming Connect.

The Welsh Pasture Project is funded by the European Agricultural Fund for Rural Development and the Welsh Government. The Farming Connect Knowledge Transfer Programme and Advisory Service are delivered by Menter a Busnes on behalf of Welsh Government. Environment Systems has been working with Menter a Busnes and IBERS (Institute of Biological, Environmental and Rural Sciences)at Aberystwyth University. The project featured on the BBC’s S4C programme, Ffermio in July.