Geospatial intelligence for the agricultural sector

To develop sustainable business models and tackle climate change, the importance of data is indisputable. Complying with rapidly changing non-financial regulations and guidelines increasingly requires companies in many industries to consider geospatial data to quantify their impact on the environment. Geospatial technology helps drive impactful change since it is a scalable innovative solution that supports companies with their strategy, decision-making processes and reporting requirements. Using the agricultural sector as a use case, we will show the potential of this technology to measure greenhouse gas emissions from deforestation.

Agriculture is vital to secure food production and to help elevate populations out of poverty. Next to logging, fires, mining, transportation and urban development, agricultural crop expansion and cattle farming have driven deforestation in the past centuries. It is estimated that mature trees can store approximately 20 kilogrammes of carbon dioxide (CO_²) per year. Deforestation therefore is one of the most significant human-imposed contributors to greenhouse gas (GHG) emissions. It is estimated that the agricultural industry is responsible for roughly one-fifth of the total worldwide GHG emissions.

What is geospatial intelligence?

Geospatial intelligence analyses image and spatial data to depict, measure and visualise geographic features (e.g., glaciers, forests or a plantation) and activities affecting the environment (e.g., deforestation, a new infrastructure project or receding glaciers). To comply with non-financial reporting regulations and guidelines, geospatial data combined with image data are becoming crucial, and even mandatory, in certain sectors to monitor environmental impact. The European Union mandates the inclusion of geospatial data for certain raw materials such as coffee, cocoa, timber, cattle and palm oil to name a few.

• Geospatial data (points, vectors) refers to any piece of information with a spatial component which usually is expressed through coordinates describing the geolocation.
  
• Image data refers to images acquired remotely either using satellite, aerial or drone-based sensors.
  

Our approach and its potential for the agriculture industry

Using geospatial intelligence, in particular satellite remote sensing imagery, is a popular and innovative method to estimate the greenhouse gas (GHG) emissions caused by deforestation. This technique has a great potential – satellite imagery covers large areas of land, and its processing can be easily automated. The process is also relatively simple – based on multi-temporal images (e.g., taken now and five years back), it is possible to produce so-called Land-Use/Land-Cover (LULC) maps, where changes in the ecosystems can be detected. Put simply, it is possible to quantify the area which has been turned from forest to farmland over a five-year period. The geospatial intelligence techniques feature a high degree of adaptability – depending on the size of the area to be monitored, the frequency of the monitoring, accuracy required and the costs.

Advancement in sensor technology has improved dramatically over the last two decades. When choosing the suitable image data for a specified area, several interrelated aspects need to be collectively taken into consideration:

Image data becomes useful when combined with geospatial data and turned into products such as land cover or land use maps where image pixels are classified to differentiate between forested area, water bodies or urban areas. With enough spectral resolution, vegetation indices can be created and provide information about vegetation health.

Companies will benefit from satellite data for different reasons. The potential need for geospatial intelligence, particularly for agricultural and food companies, is growing due to the increasing voluntary and mandatory disclosures, and specific regulations related to biodiversity and climate.

Assessing natural systems remains complex, and with the growing number of satellite data processing companies, it can be extremely challenging for companies to decide what is worth doing inhouse and what should be outsourced and to whom. Our approach consists of three essential steps summarised below and illustrated using vanilla.

In the case of a food manufacturing company, a deep understanding of the whole supply chain is paramount. Once the value chain has been mapped and material stakeholders identified, a gap analysis is performed whereby relevant regulations are reviewed and gaps identified where geospatial intelligence could bring added value.

Traceability and transparency across the supply chain can be achieved through collecting geospatial data. Data gaps along the value chain are identified down to the farm or plantation level where the raw materials originate. For every tier, it is important to understand what kind of data is needed, what can or should be asked of suppliers, what should be collected from direct sources and what can be estimated using geospatial intelligence.

At the farm or plantation level, polygon data - outlining boundaries - can be collected by either walking the perimeter or drawing the borders on a device. An alternative would be to collect point-based data overlaid with crop maps to estimate production areas. Looking at vanilla, polygons are usually collected, but with an ingredient such as palm oil, the mill location can be overlaid with palm density maps since individual palm trees are easier to identify using satellites. Using publicly available concession maps and linking them to mills with a proximity algorithm is also possible.

Refineries are usually recorded as points as well. Such data is either collected by the company itself, suppliers or estimated using open access data. Collected data can be complemented with other available open access data to increase supply chain transparency and understanding. With NGOs and consumers pushing for increased transparency in the palm industry for example, refinery and mill data is increasingly becoming publicly available.

Vanilla farm polygons are mapped to reflect the perimeter of the plot and can be overlaid first with LULC maps to estimate when the farm was established. Looking at forest layers and deforestation alerts can detect degradation. Plantation-related data is attributed to the relevant farmers and village locations, where the crop is collected. Villages are localised in the form of points. Cured beans processing facilities are also localised in the form of points and flows are connected to the village market locations. Finally, factory locations where the extraction takes place are added to the database.

Once the data is collected, stored and automatic update requirements are set up, monitoring systems can be built. These systems are important for quality, optimisation and compliance. As of 2024, the EU deforestation-free supply chains regulation will require companies and EU member states to have systems in place ensuring the import of deforestation-free commodities. To prove that a commodity is deforestation-free, companies need to have their supply chain data up to date with plantation traceability and a proper deforestation monitoring system using the best available data.

Having commodity volume flows location-dependant is beneficial because it allows for geospatial analyses, such identifying risks and opportunities. Different scenarios, such as climate specific, can be modelled at various levels of the supply chain providing insight for strategic decisions. Geospatial data is also useful to describe and monitor vegetation health and to distinguish vegetation types, e.g., crop types. Knowledge of crop types and their location can support farmers with their access to finance and insurance. Field verification and measurement is often not scalable and costly. Therefore, less developed countries often struggle to get financial support to protect their crops. Geospatial intelligence can support sustainable financing by analysing, predicting production and monitoring hazardous effects on production.

At the farm level, encroachment in protected areas where crops should not be sourced can be monitored to trigger supplier engagement or field verification. Furthermore, climatic scenarios can be modelled, for example by overlaying soil maps, precipitation patterns and how these might change, to identify potential risks for future production. At the factory level, estimated commodity flows can be combined with infrastructure data to identify optical locations to open a new facility.

Companies look for ways to reduce negative impact and set up systems that effectively measure change. Similarly, they may wish to measure positive impact achieved through value chain interventions. A robust geospatial database allows for the quantification of GHG emissions along the value chain. Carbon liability can be estimated using LULC maps and carbon storage values for above and below ground biomass. Such modelling can be finetuned with additional transparency achieved through location-based data. Geospatial analyses allow impact measurement and accurate reporting on environmental indicators

At the farm level geospatial intelligence can be used to estimate vegetation health and biomass changes over time. If trees are cut down to plant vanilla, there will be a negative impact on the biomass resulting in higher carbon emissions. On the other hand, the conversion of a monoculture vanilla plantation into an agroforestry system will result in a bigger storage of carbon through time. Geospatial intelligence can be used to verify the plantation of new plants or trees, and the protection of trees in a scalable manner, providing benefits to farmers and communities.

Outlook

As the technology advances, the amount of available data continues to increase thereby making its management, integration and interpretation challenging. The changing regulatory landscape and often inconsistent data quality and availability need to be considered, especially for long-term planning. Taking your company’s overall sustainability strategy and applicable regulatory requirements into consideration, Deloitte can support your company to develop your geospatial data strategy. This includes operational aspects such as: What type of data is needed? Should this data be collected internally, externally or estimated through open access data? What is the best way to collect, store and keep the database up to date? Also technical considerations: process setup, data availability and the selection of a suitable provider or platform, need to be considered. With our holistic sustainability services ranging from materiality assessments, sustainability reporting, regulatory compliance and assurance to sustainable supply chain, circularity and green financing, we can support your company to integrate geospatial intelligence into your strategy and reporting framework, as well as to address your broader sustainability transition needs.

In the coming months we will publish a series of articles sharing our perspective and experience on geospatial intelligence and its applications in other sectors including the financial services and fashion industry. Please reach out to us if you are interested in getting more information.