Quantifying the effect of uncertainty in soil moisture characteristics on plant growth using a crop simulation model Conor Lawless a, * , Mikhail A. Semenov a , Pete D. Jamieson b a Biomathematics & Bioinformatics, Rothamsted Research, Harpenden, Herts AL5 2JQ, UK b Crop & Food Research, Private Bag 4704, Christchurch, New Zealand Received 26 May 2007; received in revised form 29 October 2007; accepted 1 November 2007 Abstract For a given site, most readily available soil information is descriptive, qualitative and classifies soils according to origin, texture, colour and chemistry. Precise prediction of plant responses to lack of water or nutrients requires a quantitative description of moisture holding and release characteristics of soil that are often estimated from these descriptions rather than measured. In this paper we analyse the effect of uncertainty in soil description on the precision of simulated crop growth and development. This analysis suggests that accurate yield estimates depend on site-specific measurements. We demonstrate the significance of this imprecision by calculating the expected range of grain yields from a range of soils representing arable land in the UK using UK weathers and a realistic range of crop N-managements. A maximum grain yield range of 7 t/ha was found, representing a significant level of uncertainty. We conclude that since simulated yield can be so sensitive to the description of soil hydraulic properties, quantitative soil moisture characteristic measurements should be made in order to analyse plant growth in response to its environment. # 2007 Elsevier B.V. All rights reserved. Keywords: Water and nutrient availability; Crop simulation; Sirius model; Pedotransfer functions; Moisture retention and release 1. Introduction Britain has a mild climate with generally plenty of rainy days and low evaporative demand. Such conditions have led to a perception among many researchers that crop production is seldom limited by water supply (French and Legg (1979) and Penman (1971)). It seems likely that water availability should be plentiful in such a climate; nevertheless, drought responses have been observed and measured for wheat grown in sandy soils in the UK (Foulkes et al., 2001, 2002). The areas where experiments supporting these two conclusions were carried out are geographically close, and have very much the same climate. Hence the soil and its moisture retention characteristics and capacities are the main cause of the differences. However, simply classifying a soil as clay or sand is unlikely to be sufficient to quantify or predict the likely reduction in yield associated with drought. Similarly, Canterbury in New Zealand, where most of that country’s wheat is grown, is an alluvial plain with soils of varying thickness over gravel beds that are about 70% stones with soil between them. The major source of variation of root zone available water holding capacity in those soils is the depth of overlying soil which may vary from 15 cm to several metres. These variations lead to differences in optimum management (different optimum irrigation frequencies for example). If these optimum manage- ments are not implemented, yield can vary substantially across soil types. 1.1. Crop model Crop simulation models (the Sirius wheat simulation model is used as an example in this case: (Jamieson and Semenov, 2000; Jamieson et al., 1998b; Lawless et al., 2005) require several types of quantitative inputs in order to simulate the daily dynamics of crop growth and development. These inputs are: (1) daily weather including minimum and maximum tempera- tures, rainfall and radiation; (2) applied management; i.e., N and water added to the crop; www.elsevier.com/locate/fcr Available online at www.sciencedirect.com Field Crops Research 106 (2008) 138–147 * Corresponding author. Present address: Henry Wellcome Laboratory for Biogerontology Research, Newcastle General Hospital, Newcastle-Upon-Tyne NE2 4AT, UK. E-mail addresses: conor.lawless@ncl.ac.uk (C. Lawless), mikhail.semenov@bbsrc.ac.uk (M.A. Semenov), jamiesonp@crop.cri.nz (P.D. Jamieson). 0378-4290/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.fcr.2007.11.004