Configuration of the Deep Neural Network Hyperparameters for the Hypsometric Modeling of the Guazuma crinita Mart. in the Peruvian Amazon

Soil characteristics and allometric models for biometric characteristics and nutrient amounts for high yielding "Bolaina" (Guazuma crinita) trees

The Peruvian amazon is very diverse in native forestry species, the Guazuma crinita "Bolaina" being one of the most planted species in the country; however, little or no information about soil requirements and nutrient demands is known. The objective of this work was to assess the general conditions of soil fertility, biomass and macro- and micronutrient amounts in high-productivity Guazuma crinita plantations. Fields of high yielding Bolaina of different ages (1-10 years) were sampled in two regions. Soil and plant samples were collected in each field and biometric measurements of fresh weight, diameter at breast height and height were performed. For soil and plant analysis, both macro- (N, P, K, Ca, Mg, S) and micronutrients (B, Cu, Fe, Mn, Zn) were determined. Finally, allometric equations were constructed for biometric and nutrient amounts. This study is the first to assess and model macro- and micronutrient amounts in the productive cycle in this species, which grows in fertile soils. In the case of biometric equations, the logarithmic and logistic models performed better. For nutrient amounts, this species followed a pattern of Ca > N > K > P > S > Mg for macronutrients and Fe > B > Mn > Zn > Cu for micronutrients. The best prediction models for nutrients were the square root and logistic models.