Methods of estimation of height of trees in Semideciduous Seasonal Forest

O estudo foi realizado na Floresta do Rosal, Guaçuí, ES, para avaliar precisão e tempo de execução de métodos de estimação de altura, em diferentes condições de terreno, com os métodos hipsômetro Vertex (1); clinômetro digital (2); estimação com auxílio de uma régua de 15 metros (3); e estimações visuais com treinamento (4) e sem treinamento (5), em três classes de altura: 1 (15,00-17,99 m); 2 (18,00-20,99 m); e 3 (> 21,00 m). Foram mensurados 124 indivíduos em terreno plano e 87 em terreno inclinado. Os dados de altura foram comparados pelo teste t, análises gráficas de resíduos e estatísticas complementares. Para avaliar o tempo, foi realizado teste t. O método 5 apresentou o pior desempenho quanto à precisão, enquanto que o melhor desempenho foi do método 4. A declividade afetou negativamente o desempenho dos equipamentos digitais e favoreceu o método 3. Houve tendência em subestimar altura com o aumento das classes. Os métodos 4 e 5 consumiram menos tempo e o 3, o maior tempo médio. O efeito da classe no tempo médio em terreno plano foi igual para todos os métodos, à exceção do 2. Em terreno inclinado, os métodos 3 e 4 não tiveram desempenhos alterados.Palavras-chave: Altura total; precisão; tempo.AbstractMethods of estimation of height of trees in Semideciduous Seasonal Forest. The research was conducted in Forest of Rosal, Guaçuí-ES, in order to evaluate accuracy and running time of five height estimating methods in different ground conditions; the methods are as follow: Vertex hypsometer (1); Digital Clinometer (2), estimation with aid of a 15 meters ruler (3), and visual estimation with training (4) and untrained (5) in three height classes: 1 (15.00 - 17.99 m), 2 (18.00 - 20.99 m), and 3 (>21.00 m). It were measured 124 individuals in flat ground and 87 individuals in sloping ground. The height data were compared by t test and residual graphic analysis and additional statistics. To evaluate time, it was improved a t test. The method 5 had the worst performance in terms of accuracy, the best performance was improved by the method 4. The slop affected negatively the digital equipment performance and favored the method 3. There was a tendency to underestimate height as classes increase. The methods 4 and 5 consumed less time and method 3 consumed the highest average time. The class effect on average time on flat ground was the same for all methods but method 2. On slope ground, performance of methods 3 and 4 weren’t altered.Keywords: Height; accuracy; time. AbstractThe research was conducted in Forest of Rosal, Guaçuí-ES, in order to evaluate accuracy and running time of five height estimating methods in different ground conditions; the methods are as follow: Vertex hypsometer (1); Digital Clinometer (2), estimation with aid of a 15 meters ruler (3), and visual estimation with training (4) and untrained (5) in three height classes: 1 (15.00 - 17.99 m), 2 (18.00 - 20.99 m), and 3 (>21.00 m). It were measured 124 individuals in flat ground and 87 individuals in sloping ground. The height data were compared by t test and residual graphic analysis and additional statistics. To evaluate time, it was improved a t test. The method 5 had the worst performance in terms of accuracy, the best performance was improved by the method 4. The slop affected negatively the digital equipment performance and favored the method 3. There was a tendency to underestimate height as classes increase. The methods 4 and 5 consumed less time and method 3 consumed the highest average time. The class effect on average time on flat ground was the same for all methods but method 2. On slope ground, performance of methods 3 and 4 weren’t altered.Keywords: Height; accuracy; time.