Farmer risk-aversion limits closure of yield and profit gaps: A study of nitrogen management in the southern Australian wheatbelt M. Monjardino a, *, T. McBeath a , J. Ouzman a , R. Llewellyn a , B. Jones b a CSIRO, Waite Precinct, Adelaide, SA 5064, Australia b Mallee Focus, PO Box 2, Manangatang, Vic. 3546, Australia ARTICLE INFO Article history: Received 1 July 2014 Received in revised form 17 February 2015 Accepted 11 April 2015 Available online Keywords: Risk Decision analysis Crop simulation Nitrogen, dryland cropping Farming systems A B ST R AC T Nitrogen (N) is the most limiting nutrient in cereal crop production and is an important requirement in closing the gap between potential and achieved water limited yield. However, N fertiliser management in broadacre cereal cropping can be risky for farmers operating in dryland regions because of variability of rainfall and price. Farmers typically respond to this situation by making risk-averse decisions that are neither yield- nor profit-maximising. Here we use a set of case-study sites across the southern Austra- lian wheatbelt to examine the risk-return profile of a range of N management options and show the extent to which the economics of N fertiliser decisions and the farmers’ attitude to risk can determine N rates in a way that limits closure of yield gaps. Using a risk-return framework that incorporates crop simula- tion response to N application, probability theory, finance techniques, and risk-aversion analysis, we were able to better demonstrate how farmers might select N management practices that manage the trade- off between maximising economic net return and risk exposure using a risk-aversion analysis across four case-study sites. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction Agricultural practices that close yield gaps between actual and potential crop yields are receiving increased global attention in an effort to improve food security (Abeledo et al., 2008; Mueller et al., 2012; Van Ittersum et al., 2013). Reducing this gap is often achieved through the intensification of agronomic management and reduc- tion of important constraints to water-limited crop performance such as nutrition, weeds, disease and pests (Hochman et al., 2012; Lobell et al., 2009; Van Ittersum et al., 2013). However, the maximisation of water-limited yield does not necessarily result in the maximisation of profit or the preferred level of exposure to risk for dryland crop- ping farmers, particularly where large fertiliser costs are involved (Lobell, 2007; Lobell et al., 2009). In Australia, farmers in dryland cropping regions need to select fertiliser rates for their cereal crops in the face of high variability of seasons (CSIRO, 2010), as well as grain and input prices (ABARE, 2010; FAO, 2010; Price, 2009). Deciding on the best rate of fertil- iser is influenced by many factors, including grain prices, fertiliser prices, farm logistics, available finance and the crop potential yield, which is in turn controlled by available water, soil fertility, weed and pest burdens, and land use history (Angus, 2001; Chen et al., 2008; McDonald, 1989). The inherent riskiness of grain production is often high (Hayman et al., 2010; Sadras, 2002) and a two-fold in- crease in variance in wheat revenue over the past two decades has been recorded in some major Australian cropping regions (Kingwell, 2011), partly in consequence of climate-related yield variance and partly due to system change towards continuous cropping (e.g. less fallow/pasture) (Llewellyn and D’Emden, 2012). The most commonly limiting nutrient and largest fertiliser cost for Australian farmers is nitrogen (N) (Chen et al., 2008). In low- rainfall environments, such as the wheatbelt, there is an N-driven trade-off between yield per unit N and yield per unit water use (Sadras and Rodriguez, 2010). The yield-N response curve (dimin- ishing returns) means that the common practice of applying low N rates can be interpreted as an approach that maximises N use ef- ficiency (NUE), at the expense of achieving their full water-limited yield potential. As a result, N has been described as a risk-increasing input, a theory with resonance worldwide (Broun, 2007; Just and Pope, 1979; Leathers and Quiggin, 1991; Lobell, 2007; McDonald, 1989; Picazo-Tadeo and Wall, 2011; Quiggin and Anderson, 1979; Rajsic et al., 2009; Roosen and Hennessy, 2003; Russell, 1968; Sadras, 2002; Sadras and Richards, 2014; Van Herwaarden et al., 1998). Ex- ceptions are higher rainfall regions with high yield potential where over-application of N (i.e. more than necessary to attain a given yield target) has been associated with farmer risk-aversion (Gandorfer et al., 2011; Rajsic and Weersink, 2008; Rajsic et al., 2009). Strategies and tools have been successfully developed to manage the riskiness of N fertiliser decisions (Hochman and Carberry, 2011). Nevertheless, in the highly variable dryland environments of * Corresponding author. Tel.: +61 8 8303 8413; fax: +61 8 8303 8446. E-mail address: marta.monjardino@csiro.au (M. Monjardino). http://dx.doi.org/10.1016/j.agsy.2015.04.006 0308-521X/© 2015 Elsevier Ltd. All rights reserved. Agricultural Systems 137 (2015) 108–118 Contents lists available at ScienceDirect Agricultural Systems journal homepage: www.elsevier.com/locate/agsy