The biological efficiency and profitability of pasture-based systems is optimised by matching pasture supply with the nutritional demands of the herd across the year. Despite the importance of grazed pasture to the profitability of pasture-based systems, the use of purchased (i.e., non-pasture) feeds (supplementary feeds) has increased considerably over the last two decades within the New Zealand (NZ) dairy industry. Supplementary feeds may be offered to maintain or increase dry matter intake (DMI) and milk production during periods when pasture supply is insufficient to meet herd demands (i.e., a pasture deficit). However, published results indicate that, despite greater milk production and gross farm revenue (GFR), imported supplementary feed is not associated with an increase in profitability. In addition, the intensification of grazing systems through concurrent increases in stocking rate (SR) have been associated with poorer environmental outcomes, such as reduced water quality and increased greenhouse gas (GHG) emissions. The primary objective of my Masters project was to investigate the biophysical, economic, and environmental effects of removing imported supplementary feed from a typical pasture-based system in the upper North Island of New Zealand. This was conducted over three years in Northland, New Zealand (35⁰56’39”S 173⁰50’34”E), using a quantitative case-study approach that compared three pasture-based dairy farming treatments differing in SR and the nature of feed supply. As a Control treatment, imported supplementary feed was offered as palm kernel expeller (PKE) during periods of pasture deficit (~ 550kg DM/cow/yr; ~ 10% of the cows’ diet; PKE treatment). This was compared with two alternative systems to remove the need for PKE: 1) reduced SR (Pasture treatment); or, 2) growing potentially high yielding forage crops on the dairy platform and maintaining SR (Cropping treatment). I used the nutrient budgeting software OVERSEER to model treatment effects on nitrogen (N) leaching, phosphorous (P) loss, and GHG emissions. In addition, I conducted a Monte Carlo analysis to investigate economic performance of all three treatments and how it varied over a range of key input prices. On average, despite no significant difference in milk production, operating profit tended to be lower in the Cropping treatment relative to the PKE treatment. In addition, after accounting for variability in market prices, the Cropping treatment returned a lower operating profit at every probability level relative to the PKE treatment and was, therefore, an inferior system for a profit-focused decision maker. Milk production and associated GFR were lower in the Pasture treatment compared with the PKE treatment; however, there was no effect of treatment on operating profit at average market prices. When accounting for potential market price variability, the PKE treatment returned a greater operating profit in 70% of scenarios. As a result, the PKE treatment would likely provide a preferable system for a profit-focused decision maker, with low to moderate risk aversion. However, the relative profitability of treatments was highly dependent on the marginal milk production response (MMPR) to supplementary feed. Consistently large MMPRs to supplementary feed were achieved in the current study: 30 to 50% greater than published farm systems experiments. A reduction in the MMPR to supplementary feed by 10% eroded the profit advantage of the PKE treatment. In addition, despite no treatment effect on N leaching, GHG emissions tended to be lower in the Pasture treatment relative to the PKE treatment. As a result, a decision maker, paying or receiving incentives to reduce environmental externalities would likely be indifferent between the PKE and Pasture treatments, even with the large MMPR to supplementary feed achieved in the current study. Depending on the potential response to supplementary feed, profitability may be maintained with the removal of imported supplementary feeds from a pasture-based system by decreasing SR.