Temperature

Can Space Technology Solve Brazil's Agricultural Challenges? Unveiling the Impact of Land Surface Temperature on Crop Yields

January 25, 2024

-

Rosa Schmidt

Can Space Technology Solve Brazil's Agricultural Challenges? Unveiling the Impact of Land Surface Temperature on Crop Yields

In the dynamic world of agriculture, where success relies on innovation and adaptability, global forces like climate change, globalization, and deindustrialization are reshaping the industry. Adaptation requires gathering data to comprehend the challenges you face. While having data is crucial, the ability to analyze and understand it is equally important.

The amount of data available in agriculture is continuing to increase, as is the capability and capacity to effectively analyze it.  Drastic progress has been made in how precise application of inputs has become and modeling of crop growth, yield, pest and disease forecasting continues to be more and more accurate.

Why are current technologies insufficient?

As with all predictive models, what you get out is only as good as what you put in. Current predictive crop models have three key parameters of precipitation, temperature, and solar radiation. The data for these parameters come from a network of meteorological stations, which provide sufficient accuracy at the regional level, but not at the field level due to their limited spatial coverage.  

The spatial density of these weather stations varies from country to country, with some being very sparsely distributed. Differences in the surface of an area can have a major impact on how we predict temperatures. This is because the uneven surfaces affect how the air temperature is distributed.

In-field sensors have become a lot simpler to use and have come down in cost, which has meant that more in-field data can be generated, but they still lack spatial coverage and, as with weather stations, require a high level of monitoring and maintenance to ensure consistent supply of high-quality data.  

The Power of Space

With the emergence of new data analytics providers, we’ve also seen an uptake in usage of other data sources, including satellite imagery. It allows for a bird's-eye view of vast farmlands, providing insights that were once unimaginable.  

Satellite-based Normalised Difference Vegetation Index (NDVI) has provided a valuable data source to enhance predictive crop models, but it does not measure one of three key variables of precipitation, temperature and solar radiation. The challenge with near-infrared vegetation metrics, like NDVI, is that they mainly show leaf structure and greenness, but these metrics don't reveal the underlying factors affecting crop health or growing conditions.

This is where Land Surface Temperature (LST) becomes a game changer. The technology is more indicative of external factors like energy balance, weather, and soil moisture, allowing for forecasting using geophysical principles and climate change models.

Satellite derived LST enables measurements across every part of our planet, allowing us to view and analyze the temperature dynamics of different surface materials and structures. It provides us with data on one of the key parameters in predictive crop modelling, temperature, providing visibility of the variation across a field.

From LST we can derive Soil Moisture (SM), which is another key component when modelling crop yield.  Most models currently calculate this by assessing the precipitation levels, soil texture, air temperature, and in some cases in-situ soil moisture measurements.  With LST they can now start to see the variability across a field and dramatically increase the spatial resolution of their estimations.

If we consider how we have traditionally measured our own temperature, we place a thermometer in one part of our body to get a general indication. This doesn’t tell us about how the temperature is distributed across our body and which parts may be causing it to be higher or lower. Thermal cameras enable us to see the distribution of temperature across our body and gain a better understanding of what causes the fluctuations.  It is very similar to satellite LST, we can now assess the temperature variation across a field and understand where the variations are and what may cause them.

Use Case

In close collaboration with OneSoil, a digital product company that provides a platform to help farmers and agricultural companies be more profitable and sustainable, we took a closer look at agricultural production in Brazil. OneSoil uses satellite imagery for precise field boundary delineation, guiding farmers to pinpoint specific issues within the field. We wanted to understand what insights the measurement of Land Surface Temperature could offer in understanding the condition of croplands in the Western Bahia and Goiás regions.

Being the world’s biggest producer of Soybeans and sugar cane, as well as the third largest producer of Maize means that Brazil’s production has a big impact on the global food commodity market. Agriculture has rapidly expanded across Brazil over the past 50 years, with areas such as Western Bahia becoming vast agricultural plains with an ever-increasing demand for water. The climate in this region is starting to exert pressure on the crops, due to the decrease in annual precipitation over the past 40 years. Combined with high year-round temperatures, there is less water available for crops, which has led to further increase in irrigation.

There is a lack of meteorological stations in the area, but this isn’t the only difficulty, installation and articulation between them can be troublesome. Beyond that, there can be inconsistent data collection procedures and weather station maintenance in rural areas where they may lack or have poor connectivity. Purchasing, implementing and maintaining automated weather stations Brazil is an incredibly difficult due to the cost of doing so in a country of this size.  

In this image showcasing the region Bahia, we can see that there are only three local weather stations, which are all a significant distance from the fields that we are analysing.  They provide us with frequent data, but when we look at the LST, we can see a large amount of variation in temperature across the different fields. This variation in LST is as much as twenty degrees Celsius, something that can’t be detected using a weather station.

High temperatures are very detrimental on crop yield and is a parameter that has a high level of impact on the accuracy of yield estimation, as does the derived soil moisture availability.  

On September 4th, 2023, at 10:07am local time, we observed air temperature in a range of 28 to 30 Celsius degrees at the local weather stations. In the LST images we observed a range of 30 to 50 degrees Celsius across the different fields. The fields with lower observed temperatures clearly show a higher level of water availability, indicating that these fields are irrigated.  

Bahia is characterized by a diverse geographic landscape that encompasses coastal areas, extensive plains, and the semi-arid region of the Sertão. Inland, Bahia's agricultural landscape is diverse, with the cultivation of crops such as sugarcane, cocoa, soybeans, and cotton, contributing significantly to the state's economy. Additionally, the semi-arid region presents unique challenges for agriculture, with efforts focused on sustainable practices and water management.

On the other hand, we observed also Goiás as state, known for its varied terrains encompassing the Cerrado savannah to wooded regions, holds a noteworthy position in Brazil's agricultural sphere.

Goiás, a state in central Brazil, is characterized by a diverse geographic landscape that includes vast plateaus, the Brazilian Highlands, and the expansive Cerrado biome. The Cerrado is a tropical savanna known for its biodiversity and is a key part of Goiás' natural environment. In terms of agriculture, Goiás is an agricultural powerhouse, with a strong focus on soybean cultivation, cattle ranching, and the production of grains such as corn and wheat. The state's strategic location and fertile lands contribute significantly to its role as a major contributor to Brazil's agricultural output.

This is another area in Brazil that lacks the spatial coverage of weather stations to provide high resolution insights into the seasonal impact of weather on crops. Satellite LST provides us with a view of the variation in temperature across the different fields, with the irrigated areas yet again clearly visible by their cooler temperatures.

Inserting image...

On September 10th, 2023, at 10:21am local time, we observed air temperature in a range of 27 to 30 Celsius degrees at the local weather station. In the LST images we observed a range of 32 to 55 degrees Celsius across the different fields. This is a significant insight for crop modelling, as the weather stations measurements would indicate that the temperatures are not high enough to have an impact on crop yield, as they are 30 degrees or slightly below. The LST shows us a different picture, with extreme heat detected and also an indication of water availability to the crop.

We are even able to get an insight into how recently and how much irrigation has been applied by the looking at the LST. Some of the circular pivot irrigation fields have varying temperatures, which indicate different irrigation regimes across those fields and possibly different crops within them.

Increasing the spatial resolution of crop models from a regional level, down to a field level has benefits for multiple stakeholders within the food chain. Farmers have more confidence of how much products they will have to sell, enabling them to forward sell a higher percentage of their produce at a better price. Food companies are provided with a clearer picture of the production levels in different areas, from which they can source raw ingredients more efficiently. For those involved in commodity trading, the more accurate the estimation of corn and soybean production, the more efficient pricing becomes due to improved calculations of future supply.

Most importantly, improving the accuracy of crop yield prediction will lead to higher levels of food security across the world.

constellr LST brings a new level of spatial resolution and accuracy to the world of crop modelling.  Providing data on the key parameters of temperature and soil moisture at the sub-field level. This data is scalable, accurate and reliable, helping us to understand how each field behaves and manage it during these times of uncertainty.

Through strategic partnerships with providers of Crop Modelling, Data Analytics, and Digital Agriculture services, we actively develop LST based solutions. The existing satellite LST faces limitations in providing comprehensive data due to cloud cover constraints. However, the launch of new satellites with enhanced resolution and accuracy presents an opportunity to overcome these challenges. By integrating this advanced satellite data with information from weather stations, we aim to significantly enhance the precision of yield prediction models.

We firmly believe that the data derived from our satellites can play a pivotal role in addressing this challenge. Our goal is to assist farmers in identifying optimal management practices, leveraging innovative solutions to tackle the evolving challenges of tomorrow.

Curious to learn more?  

To request specific information about our technology please fill out our form via https://www.constellr.com/contact and we come back to you within 24h.