Author
 |
WILES, LORI |
|
BRODAHL, MARY |
|
Submitted to: Weed Science
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 7/3/2004 Publication Date: 11/1/2004 Citation: Wiles, L., Brodahl, M.K. 2004. Exploratory Data Analysis to Identify Factors Influencing Spatial Distributions of Weed Seed Banks. Weed Science. Volume 52:936-947. Interpretive Summary: The bank of weed seeds within corn fields is patchy so herbicide use may be reduced by targeting control for the patches, but mapping a seed bank to target control is expensive. Mapping may be more cost-effective if we knew what determines size and shape of patches and how size and shape change over time. This may best be done by comparing the distributions of seed banks of different weeds in multiple fields but the best way to analyze the data is not clear. With an untried, non-traditional statistical method, we identified factors that influenced patch size and shape and also possible mechanisms of change for seed banks of multiple weed species in 8 corn fields. Seed bank density, type of irrigation, how long seeds survive in the soil, adaptations for more extensive natural dispersal, seed size, species and soil texture determined how large and oblong patches were. How patches change over time is likely influenced the most by whether seeds are spread by wind or tillage, or are just dropped around the parent plant. We now have information to predict what patches may be like in a field and consequently, may be able to recommend ways to sample the seed bank to reduce the cost of mapping. The statistical method described complex interactions among factors controlling patch size and shape and we think it will valuable for studying the spatial dynamics of both seed banks and weed populations. Technical Abstract: Comparing distributions of different species in multiple fields will help us understand the spatial dynamics of weed seed banks, but analyzing observational data requires non-traditional statistical methods. We used classification and regression tree analysis (CART) to investigate factors that influence the spatial distributions of seed banks. CART is a method for both explaining and predicting variation in a response variable from a set of possible explanatory variables because the structure of a predictive tree model clearly depicts relationships in the data including complex interactions. We used CART to predict pattern of ranges of spatial dependence for different directions of 34 seed banks in eight corn fields. Explanatory variables were field and species attributes and seed bank density. Density and type of irrigation explained the most variation in pattern. Long ranges were associated with large seed banks and stronger anisotropy with furrow than center pivot irrigation. Pattern was also explained by seed size and longevity, adaptations for natural dispersal, species, soil texture and whether the weed was a grass or broadleaf. Significance of these factors depended on density or type of irrigation, and some patterns were predicted for more than one combination of factors. Dispersal was identified as a primary process of spatial dynamics and direction of maximum continuity and anisotropy varied for seed spread by wind, tillage or natural dispersal. Demographic characteristics and density may have been more important in this than previous research because interactions were modeled. Lack of data will be the greatest obstacle to using comparative studies and CART to understand the spatial dynamics of weed seed banks. |
