Spatial variation of crude protein content along a climatic regional gradient. Possible incidence of grazing

GONZALO IRISARRI; MARIANO OYARZABAL; SONIA CHIFLET; P. ARCURI; FEDERICO PAGNANINI; MARTÍN OESTERHELD; MARCOS TEXEIRA

Termas de Catillo, Parral

Congreso; XLII Congreso Anual de la Sociedad Chilena de Producción Animal; 2017

Sociedad Chilena de Producción Animal

IntroductionIn grazing areas of southern South America, beef production has an optimum type association with mean annual precipitation -MAP (Irisarri et al. 2014). A possible explanation for this pattern is that along the MAP gradient there are changes in plant quality, either associated with environmental controls (Epstein et al. 2002) or replacement of native covers by cultivated ones (Baldi and Paruelo 2008). Moreover, grazing intensity, strongly associated with livestock density, may also regulate forage quality intake (Pavlu et al. 2006). However, the relative effect of grazing and environment on forage quality among different functional types remains poorly understood.The general objective of this workwas to understand the spatial variation of one of the most important forage quality attributes, crude protein content, across a wide environmental gradient. The specific objectives were: (1) to describe crude protein content spatial association with MAPand mean annual temperature (MAT), and (2) describe the impact of animal husbandry on crude protein content across different plant functional types.Materials and MethodsData: We compiled 4145 values of crude protein content across a wide environmental gradient from two main sources of estimation: NIRS method (n: 1423) and Van Soest method (n: 2722). For each data point we obtained its MAP and MAT from worldClim database (www.worldclim.org/version1). Analysis: Objective 1:we performed quantile regression analysis where the dependent variable was the crude protein content and the independent variables were MAT or MAP. We analyzed the 25th and the 75th quantiles. For each quantile, we performed two types of linear regressions, simple and quadratic. We chose the best models through the Akaike criteria. Objective 2: we performed an ANOVA test where the dependent variable was the crude protein content and the independent factors were species functional type (C3, C4, legume) the origin (native or cultivated/exotic) and the estimation source (NIRS or Van Soest method). The last factor represents and direct estimation of the actual quality perceived by cattle during grazing (NIRS) vs forage quality offered for consumption (Van Soest).ResultsCrude protein content spatial variation was best explained by a positive linear association with MAP, for both quantiles. Among quantiles, there was a 3-fold difference between slopes, 25thquantile: 0,035 mg.g-1.mm-1, and 75thquantile: 0,10 mg.g-1.mm-1. However, the best model for MAT, showed an optimum type association (Fig. 1). Crude protein content reached its maximum at 9,5°C for the 25thquantile and 15,8°C for the 75th quantile.Mean crude protein content reached its highest value within the cultivated legume in vegetation class, while the lowest was for the C4 cultivated in vegetation class (Fig 2). Cultivated legumes and C3 had a higher crude protein content in the offered vegetation than in the consumed one (intake), suggesting a higher grazing pressure among these types of covers. However, the opposite was observed for C4.ConclusionsWe partially confirmed an optimum type association between forage quality (crude protein content) and environmental controls (MAT). Grazing pressure seems to represent an important forage quality determinant depending on the plant functional type (C3 and legumes).