ACTA UNIVERSITATIS AGRICULTURAE ET SILVICULTURAE MENDELIANAE BRUNENSIS SBORNÍK MENDELOVY ZEMĚDĚLSKÉ A LESNICKÉ UNIVERZITY V BRNĚ Volume LVI 28 Number 2, 2008 217 VALIDATION OF POTATO SIMULATION MODEL M. Šťastná, P. Oppeltová, J. Dufková Received: February 1, 2008 Abstract ŠŤASTNÁ, M., OPPELTOVÁ, P. DUFKOVÁ, J.: Validation of potato simulation model in the Czech Republic. Acta univ. agric. et silvic. Mendel. Brun., 2008, LVI, No. 2, pp. 217–226 Aim of the study focused on evaluation of SUBSTOR – Potato model and its utilization in potato grows management in the Czech Republic. The experimental field used for the model evaluation was located in Žabčice – South Moravia region with altitudes of 179 meters above the sea level. Tuber yield served as reference for the model evaluation. Nine years experimental data set (1994–2002) was used for the model validation. Rosara cultivar represented very early growing potato, Karin cultivar depic- ted early growing potato (Solanum tuberosum L.) in the experiment. Comparison between observed and simulated tuber yields presented the evaluation process of SUBSTOR – Potato model. Tuber yields simulated by the model showed excellent accuracy (R 2 = 0.97) for Rosara cultivar, but only for four of tested years (1997, 1998, 1999 and 2002). Karin cultivar matched lower value (R 2 = 0.43). The model tended to underestimate the tuber yield for non seasonable conditions (i.e. dry years – low amount of precipitation and its disordered distribution during the growing season or higher mean air tempera- ture) and showed the sensitivity to selected cultivars. Study proved SUBSTOR – Potato model as sui- table for utilization in potato management; however, potential differences might be expected while using the model under extreme weather conditions. evaluation, program, sensitivity, Solanum tuberosum L., SUBSTOR, weather conditions Potato is the only important tuber crop in the EU. Almost 2% of the arable land area in the EU is used for potato production (Wolf, 2002). The growing de- mand for potato – as both a fresh and processed food – and a decreasing availability of land for area ex- pansion mean that yields will have to be improved through some combination of germplasm enhance- ment, better crop protection and more efficient and productive management systems (Bowen et al., 1999). The average yield of 15.9 t . ha −1 currently es- timated at the global level is much below the yields of 30–50 t . ha −1 commonly obtained across a range of environments and management systems, so it would seem that there is considerable scope for improve- ment (Allen and Scott, 1980; 1992). Development and innovative research push scientists active in po- tato research to use more and more sophisticated computer programs to evaluate complex situations that cannot be analyzed directly with other means. The progress of dynamic crop growth models, which started more than thirty years ago, conside- rably improved analytic solution of problems in crop sciences. Those models are being used increasingly to study complex dynamic systems. Through the use of mathematical equations, researchers are better able to understand the development and improve- ment of these dynamic systems. One of the main ad- vantages of crop model application is the possibility to use them under various weather and soil condi- tions and under different environment in different regions of the world, which is not usually possible when models based on the statistical analysis are used. One of the important preconditions of the ap- plication of dynamic models is the evaluation of the model reliability in reproducing the real world processes at the given place and time (Addiscot et al., 1995; Penning de Vries, 1977; Thornton et al., 1991). The processes of evaluation of any crop model are relatively long and difficult because they require the collection of large data sets including weather, soil, crop and field management data over exten- sive time periods. The evaluation of the models nor- mally includes their validation on independent data sets i.e. defining the usefulness and relevance of