A Deeper Dive into Crop Water Stress Index: Determining Wet and Dry Thresholds

Jun 15, 20232 min read

Building on my previous discussions about the complexities of VPD measurement (link to previous post here) and the limitations of using raw canopy temperature and VPD data (link to previous post here), we explored the concept of the Crop Water Stress Index (CWSI) for informed decision-making.

The Importance of Thresholds

Recall that CWSI relies on establishing wet (non-stressed) and dry (stressed) ΔT (canopy-to-air temperature difference) thresholds specific to the plant species being monitored. These thresholds are then used to compare measured ΔT values (ΔTm) and assess plant water stress.

Calculating Wet and Dry Thresholds

There are two approaches to calculate these thresholds:

Simple Empirical Method: This method offers an easier but less accurate approach based on empirical relationships and assumptions.
Complex Modeling Approach: This method provides a more accurate but intricate approach using mathematical models.

We'll delve into the simpler method here:

Wet Threshold (ΔTl)

The empirical approach estimates ΔTl (temperature difference under non-limiting soil water availability) as a linear function of VPD: ΔTl = a - bVPD.

To determine the coefficients 'a' and 'b' in this equation:

Collect temperature and VPD data for your well-watered plants.
Plot the VPD measurements against the corresponding ΔT values.
Perform a linear regression analysis to find the best-fit line, and extract the slope ('b') and y-intercept ('a').

Dry Threshold (ΔTu)

ΔTu, the temperature difference for a non-transpiring canopy (practically a dead plant!), is typically assumed as a fixed value. Literature suggests values ranging from 2 to 5°C. However, for a more accurate estimate, you can conduct your own field measurements.

Visualizing CWSI Thresholds

You can find a graphical representation of CWSI, including wet and dry thresholds, in figure below:

Remember: While the simple method offers a starting point, the complex modeling approach provides a more robust solution. I'll explore the modeling approach in a future post.

Happy Farming!