Manejo de Megastigmus transvaalensis (Hymenoptera: Torymidae) e produtividade de Schinus terebinthifolia (Sapindales: Anacardiaceae) em área de restauração

Abstract

Efficient, sustainable, and economically viable control methods are studied for monitoring forest pests, such as the exotic wasp Megastigmus transvaalensis Hussey, 1956 (Hymenoptera: Torymidae). This wasp damages the drupes of the Brazilian pepper tree Schinus terebinthifolia Raddi, 1820 (Sapindales: Anacardiaceae) in native forests, restoration areas, and commercial crops in Brazil. Given the economic value of S. terebinthifolia and its applications in food seasoning, traditional medicine, cosmetics, bioactive compounds, the pharmaceutical industry, biopesticides, flavoring, and the recovery of degraded areas, there is a need to protect its drupes from M. transvaalensis attacks in the field. This study aimed to: i) use a physical barrier to prevent M. transvaalensis attacks and determine the parasitism rate in S. terebinthifolia drupes; ii) evaluate the drying kinetics of the drupes in the laboratory; and iii) estimate the plant's productivity in the field. Five fragments of seasonal semideciduous forest undergoing ecological restoration, with sixteen randomly selected S. terebinthifolia plants, were used. A physical barrier made of non-woven fabric (TNT) (25 cm long x 21 cm wide) was used as a treatment and applied for approximately 120 days. In the drying study, mature drupes were subjected to different temperatures (30°C and 105°C), with moisture content and mass reduction data fitted to exponential and hyperbolic models. For productivity estimation, data on diameter at breast height (DBH), total tree height (Ht), and crown area (Acrown) were collected from 16 individuals, applying Principal Component Analysis (PCA) to reduce dimensionality and predict yield. The parasitism rate was higher (60.45%) in the unprotected treatment and lower (8.06%) in the treatment protected with the fabric barrier, preserving the embryonic and epicarp integrity of the drupes. The hyperbolic model demonstrated statistical superiority, showing performance indicators for predicting both moisture content (R2Adj = 0,98; AIC = 196,52; RMSE = 0,55; X2red = 0,308) and mass reduction (R2Adj = 0,954, AIC = 254,50; RMSE = 0,706; X2red = 0,507), outperforming the exponential model. The drying kinetics showed a maximum initial rate of 0.7916% h-1 with a sharp reduction after 24h (0.1112% h-1); based on the asymptotic point (y = 1.9%), a period of 48 hours (0.0421% h-1) was established as ideal for dehydration. Finally, the productivity modeling integrated dendrometric variables and parasitism rates monitored throughout the field study, revealing that isolated variables such as DBH (p = 0.021) showed instability in the models due to multicollinearity. Thus, PCA was performed, with the first component (PC1, plant vigor) standing out as the most reliable predictor of productivity (R2Adj = 0.27; p = 0.0204), while the average parasitism rate by M. transvaalensis (P_Average) showed no significant direct linear influence on the studied sample (n = 16). In conclusion, the integration of physical exclusion methods and mathematical modeling establishes fundamental guidelines for post-harvest management and the reproducibility of forest restoration processes with S. terebinthifolia.