Insect infestations have an impact on the quality of climate reconstructions using Larix ring-width chronologies from the Tibetan plateau

Insect outbreaks are proven to harm trees, reducing their biomass accumulation due to the defoliation of leaves during the growing season. Various defoliators are responsible for the formation of extremely narrow tree rings in different tree species. In climate reconstructions based on tree-ring width, such narrow rings are often considered as noise, because their formation is not entirely caused by climatic conditions. Thus, the impact of defoliators should be removed (nullified) for unbiased climate reconstructions. This study mainly describes a statistical approach to disentangle the effects of Larch budmoth(LBM) (Zeiraphera diniana) and climate for an unbiased temperature reconstruction in the Bai-Ma mountain area in Southwest China using Larix potaninii. After applying our new approach, growth-climate relationships improved significantly (p < 0.05), especially with minimum temperature during the growing season (June-September). The resulting temperature reconstruction (BTmin) covers the period 1712-2020 CE (313 years). Our final temperature reconstruction for the Bai-Ma area showed a wiggly trend from 1712 to 1980, with some cool periods (1791-1853 and 1888-1930) and warm episodes (1724-1745, 1854-1887, 1950-1967). During the recent 50 years, BTmin reveal an alarming warming trend of up to 0.4 degrees C per decade. We also observed a strong coherency of BTmin with regional temperature series. Furthermore, we detected a differential effect of various climate modes, with the most substantial impact of the Atlantic Multidecadal Oscillation (AMO) on BTmin (p < 0.01). Interestingly, the Pacific Decadal Oscillation (PDO) showed a coupled effect along with AMO on BTmin. As revealed in this study, removing the LBM effect improved the chronology's climate sensitivity, leading to a more robust transfer function and in terms of insect infestations unbiased climate reconstructions. Thus, it is recommended to check ring-width data for any peri-odicity potentially related to insect outbreaks and correct the original data before using tree-ring chronologies for climate reconstruction.