Dairy herd production aspects of family farms in Western Amazon, Brazil

This study aimed to evaluate the production parameters of herds in 100 dairy family farms in the mesoregion of the Acre Valley, in Western Amazon, Brazil. To this end, the farms were divided into two levels of milk production. Data were collected from March to June 2016, using a 248-question semi-structured form and on-site observations. The information was recorded in SPSS (R) spreadsheets. Dairy farmers were divided into two clusters known as "high production cluster" (1,755.65 L ha(-1) yr(-1)) and "low production cluster" (492.75 L ha(-1) yr(-1)), using the K-means non-hierarchical method. Descriptive statistics was used and, with the aid of the multivariate cluster analysis, cattle ranchers were divided into the two clusters (high and low production). The results showed that the high-production cluster had larger total milk production (L milk cow(-1) day(-1)) and family income within smaller areas and using less workforce. The farmers in this group also used more ear tags for cattle identification and more technologies such as electric fence and artificial insemination at a fixed time. We concluded that family farms should improve their management and receive technical assistance to strengthen their weaknesses in dairy-cow health and reproduction systems. Moreover, milk yield in these dairy farms should be improved to increase profitability of farmers.