Skip to content

agrivoltaic_normal.md

Latest commit

 

History

History
52 lines (34 loc) · 3.53 KB

agrivoltaic_normal.md

File metadata and controls

52 lines (34 loc) · 3.53 KB
Raw

Agrivoltaic Experiment: Interpretation of Results

High-Level Summary

  • PV Panel Effect: There is strong evidence that PV Panel 2 significantly increases crop yield compared to PV Panel 1.
  • Crop Effect: There is no significant difference in yield between Crop 1 and Crop 2.
  • Interaction: There is no evidence of a significant interaction effect. This means the benefit of PV Panel 2 is consistent across both crop types; one crop doesn't uniquely prefer a specific panel.
  • Field Variation: The largest source of variation in your experiment comes from the spatial location (Blocks), which justifies including it in your model.

Detailed Interpretation of Parameters

Each parameter's mean is its estimated value, and the val2.5pc to val97.5pc range forms the 95% Credible Interval. This is the range where we are 95% certain the true parameter value lies. The key is to see if this interval includes zero.

alpha (Intercept)

  • Mean: 0.463
  • Interpretation: This is the baseline yield for your reference group, which is Crop 1 under PV Panel 1. The estimated mean yield is 0.463 units (e.g., t/ha).
  • Significance: The 95% Credible Interval is [-2.309, 4.351]. Since this is very wide and includes 0, the baseline yield isn't estimated with much precision.

beta.pv[2] (Effect of PV Panel 2)

  • Mean: 3.425
  • Interpretation: This is the most important result. On average, switching from PV Panel 1 to PV Panel 2 increases the yield by 3.425 units.
  • Significance: The 95% Credible Interval is [1.33, 6.405]. This interval is entirely positive and does not contain 0. This provides strong evidence that the effect is real and that PV Panel 2 is significantly better for yield than Panel 1.

beta.crop[2] (Effect of Crop 2)

  • Mean: 1.992
  • Interpretation: This estimates that Crop 2 yields 1.992 units more than Crop 1.
  • Significance: The 95% Credible Interval is [-0.3375, 4.897]. This interval contains 0. Therefore, you cannot conclude that there is a real difference between the yields of Crop 1 and Crop 2. The observed difference could just be due to random chance.

beta.inter[...] (Interaction Effects)

  • Interpretation: These terms test if a specific PV/crop combination has a special synergy. For example, does Crop 2 grow especially well under Panel 2?
  • Significance: Looking at all the beta.inter rows, every single one of their 95% Credible Intervals contains 0. This means there is no statistical evidence for any interaction effects. The model suggests the effects are simply additive (i.e., you get the crop effect + the panel effect).

Understanding the Sources of Variation (The sigma values)

These parameters estimate the standard deviation of the random effects, telling you where the "noise" in your data comes from.

  • sigma.block (Block Variation): 3.404
    • This is the largest source of variation. It confirms that there were significant yield differences across the physical blocks in your field. Including blockID in your model was crucial to account for this spatial noise.
  • sigma.mainplot (PV Row Variation): 0.071
    • This is the variability between the main plots (the rows of PV panels).
  • sigma.within (Residual Variation): 0.070
    • This is the random, unexplained plot-to-plot variability that's left over.

The fact that sigma.block is so much larger than the others reinforces that you designed the experiment well by blocking.