Cabruca agrosystem fungi and modeling statistic-computational of climatic variables in the plant × interaction microorganism
Trichoderma spp.; Climate; Growth biopromoter; Statistic
The southern region of Bahia has the most significant remnants of the Atlantic Forest and an agroforestry system known as cocoa cabruca, which has favorable characteristics for the emergence and/or growth of macrofungi, due to the hot and humid climate and significant contribution of matter. organic in the soil. This area is of importance for the prospection of fungi with food, medicinal, agricultural and economic potential. The objective of this work was to a) understand the relationship of climatic variables in the growth of plants treated with fungus of the genus Trichoderma spp; b) select the most important climatic variables in the statistical models that describe the relationship between the response variable (growth variable) and the fungus dosages; c) select the best statistical and computational models to evaluate the fungus × plant × environment interaction. Climatic data were obtained from NASA's public platform and correlated with growth data from plants treated with Trichoderma spp. through two approaches: multiple regression models and artificial intelligence via Random Forest. The climatic variables that most influence the growth of plants treated with the fungus of the genus Trichoderma spp are: FRICurt (Incidence of short-wave insolation on the daily horizontal surface (MJ.m-2.d-1)), FRILon (Thermal infrared radiative flux (MJ.m-2.d-1)), UmiRel (Daily relative humidity (%)), Tmed (Average temperature at 2 m from the ground daily (°C.d-1)), Tmax (Temperature maximum temperature at 2 m from the ground daily (°C.d-1)), Tmin (Minimum temperature at 2 m from the ground daily (°C.d-1)), Tmed (Average temperature at 2 m from the ground daily (°C.d-1)) , Veloc (Daily wind speed at 2 m above ground (m.s-1)), Prec (Precipitation). The climatic variable Veloc influenced only the seedling height and the variables related to temperature and radiative flux interfere in all growth variables. Therefore, these climatic variables are more important in the statistical models that describe the relationship of the response variable (growth variable) with the fungus dosages. Both forms of modeling (multiple regression and Random Forest) similarly describe the fungus × plant × environment interaction. Most of the climatic variables detected were coincident in both procedures. Future studies of application of the fungus Trichoderma spp. as a growth biopromoter, they must consider the climatic variables listed in this study, to optimize the plant seedling production process and obtain greater gains in vegetative growth.