Estimating aboveground herbaceous plant biomass via proxies: The confounding effects of sampling year and precipitation