Tillage intensity affects weed communities in agroecosystems - Invasive Plants: Ecological and Agricultural Aspects - page 187

mum L.), where linseed replaced lentil in the latter five years, 3) winter
wheat-fallow and 4) continuous winter wheat. The tillage treatments were
conventional, minimum, and zero tillage. Tillage treatments were organized as
randomized subplots within the main rotation plots. Weeds were counted by
species in twelve randomly placed 0.25 m 2 quadrats per subplot in early May,
in mid-June (prior to applying post-emergence in-crop herbicides) in the
spring-planted crops, and in October in winter wheat each year.
Multivariate analyses of weed community composition were conducted
using principal components analysis (PCA) and canonical discriminant analy-
sis (CDA) in statistical analysis system (SAS) [47]. Data were log trans-
formed, and species occurring in less than 10% of the plots were removed,
before input into PCA. Significance testing for weed community composi-
tional differences was accomplished using a data reduction step [48], where
significant axes from PCA were used as variables for input into CDA. CDA
represents multi-dimensional data structures in one or two dimensions, while
maximally preserving the trends in variation that are present in the data [49].
Ordination axes are extracted in order, with each axis summarizing the maxi-
mum amount of information not yet accounted for by previous axes. Thus, the
first axis is always the most important, followed by the second, the third, etc.
Biplots of weed species and tillage system, crop rotation and time are formed
and species associations can be ascertained by the direction of vectors. Vector
length indicates the relative strength of the association; the longer the vector
the stronger the association.
The 22 weed species in this Lethbridge study consisted of summer and win-
ter annuals, perennials, and volunteer crops. Over all rotations, zero tillage had
greater weed densities than either minimum or conventional tillage (Tab. 1).
Previous studies have similarly reported greater weed densities with zero than
with conventional tillage [50], but other studies have found that total weed
numbers were largely unaffected by tillage regime [51].

Table 1. Mean weed densities in winter wheat in May averaged over years as affected by tillage inten-
sity

Rotation Zero tillage Minimum tillage Conventional tillage

plants m 2

Winter wheat-oilseed rape 15 (5) 5 (1) 5 (1)

Winter wheat-linseed 5 (1) 11 (3) 7 (2)

Winter wheat-fallow 6 (2) 13 (4) 8 (1)

Continuous winter wheat 98 (32) 18 (4) 31 (12)

Tillage mean b 31 (9) A 12 (2) B 13 (3) B

a Values in parentheses represent standard error of the mean
b Tillage means followed by the same letter are not significantly different according to the LSD test at
the 5% level
Adapted from Blackshaw et al. [46]