Sammanfattning

Spatial and temporal variability of soil properties, irrigation schedules, and fertilization management play an important role in crop growth and yield in agricultural systems. However, how to improve and accurately predict crop yield in different management zones (MZs) within a large-scale field are important but rarely discussed. In this study, coefficient of variation, linear correlations, partial least squares discriminant analysis and quantile regression were used to quantify relationships between crop and soil properties, and determine key explanatory factors of yield, and to identify suitable ways to improve crop yield among MZs. At the temporal scale, coefficients of variation of crop growth parameters (LAI and biomass), and correlations of crop growth parameters with soil water content and available nitrogen varied during crop growth stages in MZs across years. Crop-soil correlations varied with time and were most pronounced in heading and filling stages when LAI and biomass approached maximum values, respectively. At the spatial scale, 22 explanatory factors belonging to four categories: soil physical properties, initial soil properties, soil water and nitrogen dynamics, and topography were compared. Results showed that among the four categories, initial soil properties had the greatest influence on yield across the whole field, whereas soil water and nitrogen dynamics were the most influential for individual MZs. Key explanatory factors of yield varied across MZs in 2017 and 2018. Although key explanatory factors varied across years within a MZ, some remained unchanged, especially soil physical properties. Based on the quantile regression results, the same key explanatory factors might appear in MZs, but the reference ranges were different. For more sustainable and efficient agricultural production, distributed integrated management, including soil amelioration before sowing, and irrigation and N fertilization scheduling during the crop growth period should be implemented in MZs based on the spatial-temporal variability of soil properties.

Nyckelord

Distributed integrated management; Management zones; Quantile regression; Spatial and temporal variability

Publicerad i

Agricultural Water Management
2021, Volym: 244, artikelnummer: 106535
Utgivare: ELSEVIER

SLU författare

• Parsons, David

  • Institutionen för norrländsk jordbruksvetenskap, Sveriges lantbruksuniversitet
    

Globala målen

SDG2 Avskaffa hunger, uppnå tryggad livsmedelsförsörjning och förbättrad nutrition samt främja ett hållbart jordbruk
SDG12 Säkerställa hållbara konsumtions- och produktionsmönster


UKÄ forskningsämne

Markvetenskap
Jordbruksvetenskap


Publikationens identifierare

DOI: https://doi.org/10.1016/j.agwat.2020.106535

Permanent länk till denna sida (URI)

https://res.slu.se/id/publ/110241