Has the tortoise scale exacerbated fire severity in Mediterranean stone pine forests?

Background: Introduction of non-native insect species and extreme wildfire events threaten terrestrial ecosystems and their services worldwide. However, the effect of invasive sap-feeding insect species outbreaks on fire severity is poorly understood, particularly regarding their effects on fire behavior and the probability of crown fire ignition. We set up two experimental designs to investigate how the alien tortoise scale Toumeyella parvicornis influenced fire behavior dynamics and canopy surface reflectance in Mediterranean Pinus pinea stands that were severely burnt in the summer of 2017. We combined Rothermel’s model for fire surface spread and Van Wagner’s crown ignition model to simulate fire behavior and employed data from the Landsat 8 collection to detect canopy wilt symptoms related to the multivoltine T. parvicornis abundance. Results: Simulating fire behavior in single-story P. pinea thinned and unthinned stands indicated that all the predicted fires were surface fires. Uncertainty analysis of the canopy fuel attribute model inputs revealed that fires in thinned stands were entirely classified as surface fires. In contrast, only 62.7% were surface fires in unthinned stands, whereas 37.3% were categorized as conditional fire types. Among the Landsat reflectance bands, only the NIR, green, and SWIR 2 were sensitive to the abundance of T. parvicornis. Based on these sensitive bands, two-band NIR-multiplied vegetation indices were significantly associated with the abundance of T. parvicornis from the fall generation onward when sooty mold consistently covered the canopy needles. Conclusion: The divergence between observed and predicted fires in pine stands highlights the need to investigate the processes and variables linked to T. parvicornis feeding activity on P. pinea trees to enhance fire behavior prediction. Therefore, understanding how insect outbreaks can modify fire behavior in pine stands is crucial for effective management at the local and landscape levels. Identifying the vegetation index based on sensitive bands represents an essential step toward the early recognition of insect outbreaks on a large spatial scale.