Linking frass and insect phenology to optimize annual forest defoliation estimation

Area of Science:

• Forestry
• Ecology
• Entomology

Background:

• Field measurements of forest defoliation are logistically challenging due to seasonal variations in insect feeding phenology.
• Incomplete or coarse temporal resolution field data lead to inaccurate estimations of annual defoliation impacts.

Purpose of the Study:

• To develop a novel approach for accurate estimation of annual forest defoliation using a weather-driven insect simulation model and field data.
• To improve the accuracy of defoliation impact assessments for ecosystem studies.

Main Methods:

• Leveraged the BioSIM insect simulation model with defoliation field data for *Choristoneura pinus* and *Lymantria dispar dispar*.
• Optimized an instar weighting parameter (w) and imputed defoliation using a cumulative density function (CDF).

Main Results:

• The second to last instar showed maximum consumption, improving estimates of annual frass and foliage loss.
• Achieved cross-validation RMSE of 77.53 kg·ha⁻¹ (frass) and 74.85 kg·ha⁻¹ (foliage) for *C. pinus*.
• Achieved cross-validation RMSE of 38.24 kg·ha⁻¹ (frass) and 47.77 kg·ha⁻¹ (foliage) for *L. dispar dispar*.

Conclusions:

• The novel approach provides more accurate estimates of annual defoliation, bridging gaps in field data.
• Enables better scaling of defoliation rates from field observations to regional levels using remote sensing data.
• Facilitates accurate ecosystem impact assessments for various insect taxa with variable feeding phenology.