Mineral Concentrations in Bovine Milk from Farms with Contrasting Grazing Management

Abstract

Thirty conventional and twenty-four organic dairy farms were divided into equal numbers within system groups: high-pasture, standard-pasture, and low-pasture groups. Milk samples were collected monthly for 12 consecutive months. Milk from high-pasture organic farms contained less fat and protein than standard- and low-pasture organic farms, but more lactose than low-pasture organic farms. Grazing, concentrate feed intake and the contribution of non-Holstein breeds were the key drivers for these changes. Milk Ca and P concentrations were lower in standard-pasture conventional farms than the other conventional groups. Milk from low-pasture organic farms contained less Ca than high- and standard-pasture organic farms, while high-pasture organic farms produced milk with the highest Sn concentration. Differences in mineral concentrations were driven by the contribution of non-Holstein breeds, feeding practices, and grazing activity; but due to their relatively low numerical differences between groups, the subsequent impact on consumers' dietary mineral intakes would be minor.

Keywords: dairy cow; grazing; heavy metals; macrominerals; milk; organic; pasture; production system; trace elements.

Figure 1

Concentrations of milk components and minerals where a significant interaction between pasture intake and month was observed. (a) Milk fat content in the conventional farms; (b) Milk protein content in the organic farms; (c) Milk Mo concentration in the organic farms. In (a,b), predicted means from the fitted mixed linear model are shown. In (c), means of the measured values are shown, while the p-values were obtained from the fitted mixed linear model based on the log(x + 1) transformed value. The error bars depict the means’ standard errors. Significant differences between groups within months are indicated with different letters (p < 0.05). CHP, conventional high-pasture feeding farms; CSP, conventional standard-pasture feeding farms; CLP, conventional low-pasture feeding farms; OHP, organic high-pasture feeding farms; OSP, organic standard-pasture feeding farms; OLP, organic low-pasture feeding farms.

Figure 2

Biplots of the conventional farm data (as analysed by multivariate redundancy analysis) illustrating relationships between breeding parameters (NHOL) and diet composition parameters (TFOR, GRA, GGCS, MS, LS, OMS, WC, HS, MBP, DRY, CER, CON, OIL, MIN, VIT) relative to the (i) milk yield (Yield) and contents of fat (fat), protein (prot), lactose (lact), somatic cell count (scc), aluminium (Al), calcium (Ca), copper (Cu), iron (Fe), iodine (I), potassium (K), magnesium (Mg), manganese (Mn), molybdenum (Mo), sodium (Na), phosphorus (P), tin (Sn), and zinc (Zn). Axis 1 explained 49.4% of the variation, and Axis 2 an additional 0.1%. The p-values for the drivers that contributed to the explained variation was as follows: NHOL (p = 0.002), MS (p = 0.002), CON (p = 0.002), GGCS (p = 0.002), OIL (p = 0.002), LS (p = 0.002), TF (p = 0.006), OMS (p = 0.016), HS (p = 0.018), MBP (p = 0.036), VIT (p = 0.038), MIN (p = 0.138), DRY (p = 0.272), GRA (p = 0.696).

Figure 3

Biplots of the organic farm data (as analysed by multivariate redundancy analysis) illustrating relationships between breeding parameters (NHOL) and diet composition parameters (TFOR, GRA, GGCS, MS, LS, OMS, WC, HS, MBP, DRY, CER, CON, OIL, MIN) parameters relative to the (i) milk yield (Yield) and contents of fat (fat), protein (prot), lactose (lact), somatic cell count (scc), aluminium (Al), calcium (Ca), copper (Cu), iron (Fe), iodine (I), potassium (K), magnesium (Mg), manganese (Mn), molybdenum (Mo), sodium (Na), phosphorus (P), tin (Sn), and zinc (Zn). Axis 1 explained 37.2% of the variation, and Axis 2 an additional 0.3%. The p-values for the drivers that contributed to the explained variation was as follows: TF (p = 0.002), NHOL (p = 0.002), MS (p = 0.032), MIN (p = 0.062), GGCS (p = 0.064), DRY (p = 0.158), MBP (p = 0.190), HS (p = 0.216), GRA (p = 0.236), OIL (p = 0.332), LS (p = 0.708), CON (p = 0.738), WC (p = 0.788).