Revealing the determinants of wheat yields in the Siberian breadbasket of Russia with Bayesian networks

Alexander V. Prishchepov*, Elena Ponkina, Zhanli Sun, Daniel Müller

*Corresponding author for this work

    Research output: Contribution to journalJournal articleResearchpeer-review

    29 Citations (Scopus)

    Abstract

    Higher crop yields are critical to satisfy the rising global food demand. Russia holds untapped potential for increasing agricultural production because current grain yields are often far below the potentially attainable yields. Western Siberia is an important breadbasket in Russia, where wheat yields fall particularly short of their potential. Our goal was to assess the determinants of yield variations among farmers in the province of Altai Krai in Western Siberia. We conducted 67 structured in-person interviews with corporate farm managers and individual farmers about the potential determinants of wheat yields and complemented these data with 149 additional observations obtained from the provincial agricultural extension service. We used Bayesian networks (BNs) to represent the relationships between the explanatory parameters and contemporary wheat yields and to examine qualitative future scenarios of future yields. The results revealed higher yields on larger farms than on medium and small farms. Our results corroborated that the application of fertilizers and herbicides and the implementation of new equipment had large positive impacts on the yields. The scenario of higher future production costs and lower precipitation resulted in a yield reduction from 7.6 dt/ha to 5.3. Overall, our results suggest that policies aimed at increasing wheat yields should concentrate on the education of farmers and encourage higher input applications, particularly for small-scale farms. Additionally, policies should address concurrent challenges, such as a higher drought frequency, through the application of new equipment, seed material and tillage practices.

    Original languageEnglish
    JournalLand Use Policy
    Volume80
    Pages (from-to)21-31
    Number of pages11
    ISSN0264-8377
    DOIs
    Publication statusPublished - 2019

    Keywords

    • Bayesian belief network
    • Food security
    • Land-use intensity
    • Russia
    • Scenario analysis
    • Wheat production
    • Yield gap

    Cite this