Hair of grazing cattle provides an integrated measure of the effects of site conditions and interannual weather variability on δ¹³C of temperature humid grassland

The carbon isotope composition (δ¹³C) of C3 ecosystems is sensitive to water availability, and provides important information for the assessment of terrestrial carbon (C) sink/source activity. Here, we report the effects of plant available soil water (PAW) on community ¹³C signatures of temperate humid grassland. The 5-year study was conducted on pastures exhibiting a large range of PAW capacity that were located on two site types: peat and mineral soils. The data set included the centennial drought year 2003, and data from wet years (2000 and 2002). Seasonal variation of PAW was modeled using PAW capacity of each pasture, precipitation inputs and evapotranspiration estimates. Community ¹³C signatures were derived from the δ¹³C of vegetation and segments of tail switch hair of cattle grown while grazing pastures. Hair ¹³C signatures provided an assimilation-weighted ¹³C signal that integrated both spatial (paddock-scale) and temporal (grazing season) variation of ¹³C signatures on a pasture. The δ¹³C of hair and vegetation increased with decreasing modeled PAW in the same way on mineral and peat soils. But, at a given PAW, the δ¹³C of hair was 2.6‰ less negative than that of vegetation, reflecting the diet-hair isotopic shift. Furthermore, the δ¹³C of hair and vegetation on peat soil pastures was 0.5‰ more negative than on pastures situated on mineral soil. This may have resulted from a ∼10 ppm CO₂ enrichment of canopy air derived from ongoing peat mineralization. Community-scale season-mean ¹³C discrimination (Δ) exhibited a saturation-type response towards season-mean modeled PAW (r² = 0.78), and ranged between 19.8‰ on soils with low PAW capacity during the drought year of 2003, and 21.4‰ on soils with high PAW capacity in a wet year. This indicated relatively small variation in season-mean assimilation-weighted p_i/p_a (0.68-0.75) between contrasting sites and years. However, this range is similar to that reported in other studies, which encompass the range from subtropical arid to humid temperate grassland. Furthermore, the tight relationship between season-mean Δ and modeled mean PAW suggests that PAW may be used as proxy for Δ.