Influence of soil properties on crop yield: a multivariate statistical approach

Abstract

A b s t r a c t. The aim of the study was to reveal the relationship between soil properties and grain yields in an East Hungarian region in regard to weather conditions. Soil pH, EC, carbonate content, soluble and exchangeable Na+, texture, organic carbon, and nutrient contents were analyzed. Yield data (maize, winter wheat, sunflower) from 10 years were standardized using calculated relative yield and yield variability. Weather conditions were characterized by the Pálfai Drought Index. Hydrological and topographical conditions were characterized by the mean altitude of plots. The ranged pedological variables were analyzed using principal component analysis with Varimax rotation. The principal component analysis showed that three principal components with eigenvalues greater than one explained more than 84% of the variability of soil properties. The multiple stepwise principal regression analysis showed that the mean relative yield was linearly correlated with all the three principal component factors (R 2 = 0.49, p < 0.01). In droughty years, the sodification, salinization, soil texture, and nutrient contents determined the yields (R 2 = 0.30, p < 0.05). In humid years, the lower topographical position, soil organic matter, and nutrient contents were the main limiting factors (R 2 = 0.40, p < 0.01). Consequently, the variables can effectively explain the yield variability together with other variables as linear combinations.

Multiple linear regression and random forest to predict and map soil properties using data from portable X-ray fluorescence spectrometer (pXRF)

“…In this sense, the use of machine learning tools may accelerate the identification of data for characterizing soils. Several methods of analyzing large amount of data of both continuous and categorical variables have been used in works of various natures, such as the stepwise multiple linear regression (SMLR) (Juhos;Szabó;Ladányi, 2015;Menezes et al, 2016;Rodrigues;Corá;Fernandes, 2012). This analysis adjusts regression models from easily obtained variables to estimate data more difficult to be acquired, in which the addition or removal of predictive variables to the model is performed based on statistical tests, generating a final equation.…”

Fruit Quality of Tomato Affected by Single and Combined Bioeffectors in Organically System

“…As a consequence of intensive agrochemical application, environmental impacts on soils are increasing parallel with loss of abundance and diversity of living organisms (Johnston, 1986;Varga et al, 2007). One of the most important requirements, therefore, is the sustainable farming practices and food production with the need to reduce the use of artificial chemicals and stimulants (Juhos, 2014;Juhos et al, 2015Juhos et al, , 2016Tariq et al, 2016). With those conditions consumers' highlighted expectation for healthy and chemical-free foods are also required.…”

Yield Prediction of some Orchard Trees Based on Soil Characteristics under Rainfed Conditions

“…understanding the variability of landscape and soil characteristics and their influences on crop productivity is a vital and critical component of the site-specific and sustainable management system and land use planning, (Juhos et al, 2015). The expectable yield or productivity capacity is useful in assessing the soil suitability for agricultural use.…”

“…Similarly, (Mbodj et al, 2004), in Oud Rmel Catchment of Tunisia, found that the most influential limiting factors were alkaline pH and the excessive amount of the soil calcium carbonate. Multiple statistical procedures have been developed to presage the crop yield, the fitness of these procedures relies on the framework and size of the database, but each method has its own limitations, (Juhos et al, 2015).…”