A Comprehensive Guide to Estimating Field Biomass Using NDVI and Direct Sampling

Introduction

 Biomass is the total mass of plant tissue within a specific area or environment. It is frequently represented in terms of dry mass, which excludes water content for uniformity of measurement. It is an essential indicator for determining the health and productivity of a cropping system or pasture in an agricultural setting.

Biomass is typically measured directly, such as by clipping and weighing plant material in a representative area, utilizing tools like a rising plate meter, or using a pasture ruler for measurements. Remote sensing technologies such as drone and satellite images, as well as NDVI (Normalized Difference Vegetation Index) imagery, may also be used. NDVI detects plant health and and can be used as a density factor when measuring biomass.

Use in Grazing systems

Biomass and dry matter estimates for pastures are important components of managing grazing systems for cattle production. Knowing the amount of available forage in a field can help farmers and ranchers make better decisions on stocking rates, grazing rotations, and overall herd management, reducing the need to purchase feed.

Traditional grazing systems often rely on set stocking rates and continuous or rotational grazing practices that do not frequently adapt to changing forage conditions. Estimating pasture biomass may not be used in this system, but ignoring this factor or estimating with low accuracy can lead to excess cost for imported feed.

AMP grazing, intensive rotational grazing, and regenerative grazing practices involve moving cattle frequently between small paddocks to mimic natural grazing patterns. This system aims to improve pasture health and productivity through rest periods that allow for regrowth. Biomass and dry matter estimates in AMP systems are more important, because they can be a key factor in determining rotation schedules.

Our Method

The way we estimate biomass combines NDVI imagery with direct field sampling techniques. NDVI is a widely used index that provides a measure of plant health by comparing the difference between near-infrared (which healthy vegetation strongly reflects) and red light (which vegetation absorbs) High NDVI values indicate healthy, dense vegetation.

Step 1: Field Sampling

Direct field sampling involves harvesting the vegetation in area of a specific size (usually inside a hoop or square frame of a set size, drying it, and then weighing, then converting that into lbs/acre. By integrating field sampling with NDVI data, we can extrapolate those measurements across larger areas with better accuracy. Steps include:

1. Determine Locations for Measurement: Select locations across the field that will represent the variability plant growth. Avoid gathering all samples from areas that look similar to get a representative understanding of the biomass across the entire pasture or field.

2. Mark Out a Sampling Area: Use a standardized tool such as a hoop or square frame to mark out the exact area for vegetation collection. Larger frames provide a more representative sample but require more effort to handle and process. 1 square foot is the industry standard, but we like using 2 square feet to be a bit more representative.

3. Collect Vegetation Samples: Clip all vegetation within the marked area to 3 inches above ground for pastures. This estimates the forage available for grazing.

4. Document Sample Location: Use point collection software or apps like Pix4Dcatch to record the location of each sample area. This georeferencing is crucial for repeat measurements, spatial analysis, finding the NDVI value of the point, and importing those points into software like Arcgis. Pix4D catch allows you to export your points to an excel file.

5. Weigh Wet Samples: Immediately after collection, weigh the clipped vegetation using a kitchen or field scale to determine the wet mass.

6. Dry the Samples: Spread the samples evenly in a drying oven at a low, consistent temperature (55-65°C) until they reach a constant weight. Drying times may vary depending on vegetation type and moisture content.

7. Weigh Dry Samples: After drying, weigh the samples again to obtain the dry mass.

8. Convert Sample Weight to lb per Acre: To extrapolate the sample weight to a larger area, use the formula: Biomass (lb/acre)= Sample Dry Weight (oz / square foot )/16) × 43,560

Step 2: Access NDVI Imagery

We use Pix4D Fields and ArcGIS to pull satellite imagery. The following are general steps for either software..

Pix4Dfields

1. Capture or Import Images: Start by capturing images using a drone equipped with a multispectral camera. NDVI satellite imagery can be downloaded directly in pix4d.

2. Create a Project: Open Pix4Dfields and create a new project. Import your images into the project. If you have an existing field boundary created for your field, you can download NDVI imagery from a satellite directly to Pix4D using the import satellite imagery button.

3. Process the Imagery: Use Pix4Dfields to process the imported images. The software will stitch the images together to create a single, detailed map of your area of interest.

4. Generate NDVI Maps: Once the images are processed, select the option to generate NDVI maps. Pix4Dfields automatically calculates NDVI values using the available spectral bands from your imagery.

5. Analyze and Export: Analyze the generated NDVI map within Pix4Dfields. Export the NDVI map for further analysis in ArcGIS. ArcGIS allows users to see the NDVI value for individual cells, which is important for model creation and interpolation.

Using Sentinel-2 Imagery on ArcGIS:

The Sentinel-2 mission provides global coverage with multispectral data. To use Sentinel-2 imagery from ArcGIS:

1. Find Sentinel-2 Imagery: Use the link provided (Sentinel-2 Imagery on ArcGIS) to access the Sentinel-2 imagery layer. You can also search for "Sentinel-2" in the ArcGIS Online search bar. Copy the URL for this dataset.

2. Add to Map: Once you find the Sentinel-2 imagery layer, add it to a new or existing map by clicking on Map, Add Data, Data from path. And enter the URL for Sentinel imagery.

3. Open the Map Viewer: With the Sentinel-2 layer added, open the map in the Map Viewer.

4. Select NDVI Raw: Use the data tool to choose the NDVI Raw processing template.

5. Visualize and Analyze: After applying the NDVI calculation, adjust the symbology settings to visualize the NDVI values effectively.

6. Import Points or Locate Points: Import or locate your sample locations and note the NDVI value for each of your sample locations.

7. Find the average NDVI: If your dataset is clipped to the size of your field (ie. imagery imported from Pix4Dfields). The mean NDVI Value will be noted in the statistics.
   

1. 
   

Interpolation and Model Development

When you have multiple pairs of NDVI and Dry Matter (DM) values and you want to find the DM value for an average NDVI value, you're essentially looking to fit these data points onto a curve or line and then find the corresponding DM value for your average NDVI. For simplicity, we'll focus on linear interpolation here, which works when the relationship between NDVI and DM can be reasonably approximated by a straight line.

Linear Interpolation

With multiple data points, you first need to determine which two points your average NDVI value falls between. This is straightforward if the NDVI values are ordered. Once you've identified the appropriate interval, you can apply the formula directly.

y = y 1 + ( x − x 1 ) ( y 2 − y 1 ) / (x 2 − x 1)  or

DMave = DMlow  + (NDVIave - NDVIlow) * (DMhigh - DMlow)/ (NDVIhigh -NDVIlow)

Note

This method assumes a linear relationship between NDVI and DM for the interval you're working with. If the relationship is more complex, or if you want to interpolate using all points simultaneously (e.g., polynomial interpolation or using a regression model), the approach would be different and might require statistical software or programming tools.

Linear Regression in Excel

Linear regression will allow you to build a prediction equation

1. Collect Data: Gather your pairs of NDVI and DM values. Collect as many as possible and add to a prediction data set over time.

2. Order data in excel

3. Select data and insert scatter plot

4. Click a point and click add trendline

5. Choose display equation

6. Create a new column and input value for NDVI from 1 to 0 with the level of detail needed.

7. Use formula from regression model to fill a prediction cell..

8. Choose the value for the average NDVI for the field.

This image is not from real data. It is a sample to show how to set up a regression model in excel.

For nonlinear relationships, polynomial regression or other non-linear models might provide a better fit but require more complex calculations.