Developing transfer functions for impact-abundance relationships in defoliating geometrid moths

The outbreak regimes of forest insects are altered by climate change. This accentuates the need for a quantitative understanding of the relationship between insect abundance and host plant impact, and for tools which permit impact monitoring across large areas. We developed three transfer functions that link ground-based larval density estimates of three severe forest pests ( Epirrita autumnata , Operophtera brumata , Agriopis aurantiaria ), field-estimated crown defoliation, and a satellite proxy of defoliation (MODIS NDVI anomaly) in subarctic mountain birch forests in northern Norway. We combined long-term larval counts from 274 stations at 18 localities (9–23 years) with a one-year (2014) survey of tree-level defoliation at 90 stations and pixel-wise NDVI anomalies (2000–2023). Transfer functions were developed using mixed-effects models. Transfer function 1 (TF1) showed a strong, saturating increase of defoliation with larval density (marginal R² = 0.76; conditional R² = 0.84), reaching ∼80–90% defoliation at high densities. Transfer function 2 (TF2) revealed a negative linear relationship (slope −0.07 NDVI anomaly per 1% defoliation; R²m = 0.34; R²c = 0.65). Transfer function 3 (TF3) identified a detection threshold at 26.4 larvae per station (95% CI 18.8–36.9): below this, NDVI anomalies were insensitive to changes in larval density; above this, anomalies declined steeply, with substantial location-specific variation (breakpoints 14.8–56.7 larvae/station). Together, these transfer functions provide operational links between established abundance monitoring and satellite-based impact mapping, define detection limits for remote sensing, and enable reconstruction and forecasting of outbreak impacts. We outline avenues to further improve sensitivity using higher-resolution sensors and repeated field calibration. • Transfer functions (TF1-3) between ground- and remote sensing-based estimates of defoliator density and canopy defoliation • TF1 describes a strong saturating increase in tree level defoliation with increasing defoliator density • TF2 describes a linear negative relationship between tree level defoliation and a remote sensing defoliation proxy • TF3 identifies a lower detection threshold in the relationship between defoliator density and the defoliation proxy • The functions provide tools to integrate field monitoring with remote sensing for operational large-scale monitoring and forecasting