Elsevier

Geoderma

Volume 140, Issue 3, 15 July 2007, Pages 235-246
Geoderma

Role of landscape and hydrologic attributes in developing and interpreting yield clusters

https://doi.org/10.1016/j.geoderma.2007.04.008Get rights and content

Abstract

Management of agricultural fields based on yield patterns may help farmers adopt environmentally friendly farming practices. Our objective was to investigate landscape and hydrologic attributes that affect spatial clusters of corn (Zea mays L.)–soybean (Glycine max L.) yields. The study was conducted at Iowa State University's northeastern research center near Nashua, Iowa, from 1993 to 1998. The yield data, normalized for annual climatic variability, were used in cluster and discriminant analysis, and the landscape and hydrologic data were overlain using ArcGIS software. Three clusters of low, medium and high categories were formed using 10 iterations with zero convergence options and satisfying the R2, pseudo F-statistic and cubic clustering criteria. The spatial clusters, however, varied greatly over space and time domain for the study period. The map overlay analysis using ArcGIS showed that high yield clusters were affected by soil and lower elevation levels in the below average precipitation year of 1994. The annual normalized subsurface drainage volume, nitrate leaching losses, soil type and topographic attributes of slope, aspect, and curvature were used in stepwise discriminant analysis to identify the variables significantly related to the clusters. Soil and topographic attributes of curvature and aspect contributed significantly in cluster formations for four of the six years at P  0.15. The results suggest that cluster and discriminant analysis can be useful for identification of soil and topographic attributes affecting corn and soybean yield patterns, which can help in delineation of management zones for site specific management practices.

Introduction

Crop yields are an outcome of the complex interaction among soil, topography, climate, and management practices and vary considerably within a field (Kanwar et al., 2005, Jaynes et al., 2005, Bakhsh and Kanwar, 2005, Reuter et al., 2005). Crop yield variability within a field may be due to intrinsic factors such as soil type, soil moisture, and nutrient potential as well as the extrinsic factors such as climate and management practices (Lamb et al., 1997, Bakhsh et al., 2000). Climatic variables are the most dominating factors that also cause temporal variability of crop yields within a field from year to year. The precipitation and temperature are the main two driving factors besides several others in affecting the water and nutrient availability to plants (Mulla and Schepers, 1997, Bakhsh et al., 2002, Baker et al., 2005).

In addition to climatic effects, the soil moisture conditions within the soil profile are also affected by the landscape attributes and soil texture to move and retain water in the soil profile. Iqbal et al. (2005) reported that soil properties vary with the topographic settings and influence the redistribution of soil water content along the slope. Machado et al. (2002) reported that water, elevation and soil texture consistently influence the grain yields. Kravchenko and Bullock (2000) also examined the effects of topographic attributes and the derived hydrologic indices on variability in soil properties and crop yields. In other words soil characteristics and topography play an important role in varying the crop yield patterns within a field or watershed due to spatial variability effects (Afyuni et al., 1993, Fiez et al., 1994, Fraisse et al., 2001). The impact of topographic attributes on crop yield becomes more important especially for the soils having subsurface drainage ‘tile’ system (Bakhsh and Kanwar, 2004).

Subsurface drainage is imperative for the soils in the midwestern parts of the United States to maintain their productivity potential by removing excess water from the root zone. The subsurface drainage systems, however, have also been reported to enhance the chemical transport from the bottom of the root zone to the edge of the field (Jaynes et al., 1999, Dinnes et al., 2002, Bakhsh and Kanwar, 2005). One such chemical, important for plant growth is nitrate-nitrogen (NO3-N) which is soluble and non-adsorbent in nature, moves freely with the soil water and exits the system with subsurface drainage water (Baker et al., 1997). Leaching of NO3-N via subsurface drainage water has caused a serious environmental concern as well as economic loss to the farming community. The USEPA (1995) has identified the agriculture sector as one of the major contributors to soil and water pollution. The upper Mississippi river basin has been reported to be exporting 39% of the N delivered to the Gulf of Mexico which is suffering from the second largest hypoxic zone in the world (Alexander et al., 1995, USGS, 2005). Several studies have linked the hypoxic zones to the NO3-N loadings from the Mississippi river basin (Rabalais et al., 2002, Randall and Vetsch, 2005, Kanwar et al., 2005). Therefore to develop sustainable farming systems that minimize NO3-N leaching and maximize crop yield necessitate studying the role of soil and landscape features in transporting NO3-N from agricultural fields.

Leaching of NO3-N to subsurface drainage water depends on its concentrations in the soil profile at the time of water percolation below the root zone. The concentrations of NO3-N in the root zone have been reported to be affected by the controllable factors of management practices in addition to the uncontrollable factors of climatic variables (Kanwar et al., 1988, Bakhsh et al., 2002, Baker et al., 2005). Dinnes et al. (2002) concluded that N dynamics in agricultural fields in humid regions is affected by a number of factors including tillage, drainage, crop type, soil organic matter content and weather conditions. The interaction of climatic variables, soil and topography has caused the spatiotemporal variability among and within the fields despite having the same management treatments. Bakhsh and Kanwar (2005) have reported spatial variability effects on NO3-N leaching losses to subsurface drainage water for the field plots under the same management practices. They further recommended that site specific management of the soils needs to be made to reduce the offsite transport of NO3-N from agricultural fields. The spatial zones need to be delineated for site specific management practices.

One approach to address the spatial variability effects of the soils and topographic attributes can be grouping of the response variables into meaningful interpretable zones and using the map overlay capability of Geographic Information System (GIS) to study the spatial relationships (Bakhsh and Kanwar, 2004). GIS is a powerful tool for determining the integrated effects of the various soil and landscape data layers with crop yield patterns (Bakhsh et al., 2000). The effects of soil and topographic attributes on yield variation can be perceived better when data layers of these attributes are overlaid (Silva and Alexandre, 2005).

Quantification of the soil and landscape effects on crop yield is essential to agricultural decision making. For example, adoption of conservation practices requires weighing the benefits to the environment with the affects on crop production under the site specific soil and climatic conditions. Malone et al. (2006, in this issue of Geoderma) quantified corn yield based on variable climate and N applications. As discussed above, soil and landscape attributes significantly affect crop yields, therefore, agricultural decision making tools should account for these effects.

Crop yield has also been considered a good indicator for delineation of stable management zones (Bakhsh et al., 2000). Cluster and discriminant analyses have been used to classify the crop yield data into meaningful groups and study the contribution of various soil and landscape attributes in discriminating these clusters, respectively. These approaches have been used in different disciplines by several researchers (Al-Sulaimi et al., 1997, Bakhsh and Kanwar, 2004, Kaspar et al., 2004, Jaynes et al., 2005, King et al., 2005). Bakhsh and Kanwar (2005) applied an integrated approach to study the offsite transport of NO3-N leaching losses by developing spatial clusters and seeking their spatial relationships with the soil and topographic attributes. They reported that spatial NO3-N leaching losses clusters were affected by the interaction of soil type and elevation levels. No study, however, has been conducted to investigate the spatial yield clusters for the soils having subsurface drainage and determining their relationships with the subsurface drainage flows, NO3-N leaching losses and the landscape attributes. The hypothesis of this study was that soil and landscape attributes can affect the crop yield patterns and have spatial relationships with the yield clusters. The specific objectives of the study were:

  • Delineate spatial zones of corn–soybean yields using cluster analysis.

  • Identify landscape and hydrologic attributes that contributed significantly in discriminating yield clusters using discriminant analysis.

  • Integrate and overlay GIS data layers of the identified landscape and hydrologic attributes on crop yield clusters to establish cause–effect relationships.

Section snippets

Materials and methods

The field experimental data on corn–soybean yields from 1993 to 1998 were collected from 36 plots, each 0.4 ha in size, at Iowa State University's north-eastern research center near Nashua, Iowa. The soils at the site are Floyd loam (fine-loamy, mixed, mesic, Aquic Hapludolls), Kenyon silty-clay loam (fine loamy, mixed mesic, Typic Hapludolls) and Readlyn loam (fine-loamy, mixed, mesic, Aquic Hapludolls) (Kanwar et al., 1997). These soils are moderately well to poorly drained, lie over loamy

Corn–soybean yields

Average corn grain yields varied greatly from 5.48 Mg ha 1 in 1993 to 9.19 Mg ha 1 in 1997 which show the effect of growing season rainfall variability because 1993 was a wet year (Table 1). The year 1993 had a rainfall of 1030 mm compared with 750, 800, 680, 750 and 980 mm for 1994, 1995, 1996, 1997, and 1998, respectively. This compares to the normal growing season rainfall of 771 mm for the study area (Voy, 1995). The amount and distribution of rainfall during the growing season are

Summary

Grouping of response variables into meaningful classes can help manage their spatial occurrence. In this study, three classes of clusters namely low, medium, and high were generated for each year using corn–soybean yield data from 1993 to 1998. Cluster formation was found to be satisfactory using the evaluation criteria of R2, pseudo F-statistics and cubic clustering criteria. Corn–soybean grain yield clusters varied greatly over the years because of changing rainfall patterns. The distribution

References (51)

  • A. Bakhsh et al.

    Spatial clusters of subsurface drainage water NO3-N leaching losses

    J. Am. Water Resour. Assoc.

    (2005)
  • A. Bakhsh et al.

    Using soil attributes and GIS for interpretation of spatial variability in yield

    Trans. ASAE

    (2000)
  • A. Bakhsh et al.

    Cropping system effects on NO3-N losses with subsurface drainage water

    Trans. ASAE

    (2002)
  • D.L. Bjorneberg et al.

    Seasonal changes in flow and nitrate-N loss from subsurface drains

    Trans. ASAE

    (1996)
  • D.L. Bjorneberg et al.

    Alternative N fertilizer management strategies effects on subsurface drain effluent and N uptake

    Appl. Eng. Agric.

    (1998)
  • A. Blackmer et al.

    Correlations between soil nitrate concentrations in late spring and corn yields in Iowa

    J. Prod. Agric.

    (1989)
  • A. Changere et al.

    Slope position and erosion effects on soil properties and corn production on a Mianian soil in central Ohio

    J. Sustain. Agric.

    (1997)
  • T.S. Colvin et al.

    Yield variability within a central Iowa field

    Trans. ASAE

    (1997)
  • A.J. Conacher et al.

    The nine unit landsurface model: an approach to pedogeomorphic research

    Geoderma

    (1977)
  • N.A.C. Cressie

    Statistics for Spatial Data

    (1993)
  • D.L. Dinnes et al.

    Nitrogen management strategies to reduce nitrate leaching in tile-drained Midwestern soils

    Agron. J.

    (2002)
  • T.E. Fiez et al.

    Winter wheat yield and grain protein across varied landscape positions

    Agron. J.

    (1994)
  • C.W. Fraisse et al.

    Delineation of site specific management zones by unsupervised classification of topographic attributes and soil electrical conductivity

    Trans. ASAE

    (2001)
  • J. Iqbal et al.

    Relationship between soil-landscape and dryland cotton lint yield

    Soil Sci. Soc. Am. J.

    (2005)
  • D.B. Jaynes et al.

    Spatiotemporal variability of corn and soybean yield

    Agron. J.

    (1997)
  • Cited by (0)

    1

    Visiting Scientist (on sabbatical leave, Associate Professor, Department of Irrigation and Drainage, University of Agriculture, Faisalabad, Pakistan).

    View full text