Anthropogenic forcing on the spatial dynamics of an agricultural weed: the case of the common sunflower

Generating an agricultural risk map based on limited ecological information: A case study using Sicyos angulatus

In this study, we propose a method for estimating the risk of agricultural damage caused by an invasive species when species-specific information is lacking. We defined the "risk" as the product of the invasion probability and the area of potentially damaged crop for production. As a case study, we estimated the risk imposed by an invasive weed, Sicyos angulatus, based on simple cellular simulations and governmental data on the area of crop that could potentially be damaged in Miyagi Prefecture, Japan. Simulation results revealed that the current distribution range was sufficiently accurate for practical purposes. Using these results and records of crop areas, we present risk maps for S. angulatus in agricultural fields. Managers will be able to use these maps to rapidly establish a management plan with minimal cost. Our approach will be valuable for establishing a management plan before or during the early stages of invasion.

Ecologically sustainable weed management: How do we get from proof-of-concept to adoption?

Weed management is a critically important activity on both agricultural and non-agricultural lands, but it is faced with a daunting set of challenges: environmental damage caused by control practices, weed resistance to herbicides, accelerated rates of weed dispersal through global trade, and greater weed impacts due to changes in climate and land use. Broad-scale use of new approaches is needed if weed management is to be successful in the coming era. We examine three approaches likely to prove useful for addressing current and future challenges from weeds: diversifying weed management strategies with multiple complementary tactics, developing crop genotypes for enhanced weed suppression, and tailoring management strategies to better accommodate variability in weed spatial distributions. In all three cases, proof-of-concept has long been demonstrated and considerable scientific innovations have been made, but uptake by farmers and land managers has been extremely limited. Impediments to employing these and other ecologically based approaches include inadequate or inappropriate government policy instruments, a lack of market mechanisms, and a paucity of social infrastructure with which to influence learning, decision-making, and actions by farmers and land managers. We offer examples of how these impediments are being addressed in different parts of the world, but note that there is no clear formula for determining which sets of policies, market mechanisms, and educational activities will be effective in various locations. Implementing new approaches for weed management will require multidisciplinary teams comprised of scientists, engineers, economists, sociologists, educators, farmers, land managers, industry personnel, policy makers, and others willing to focus on weeds within whole farming systems and land management units.

The effects of temporally variable dispersal and landscape structure on invasive species spread

Many invasive species are too widespread to realistically eradicate. For such species, a viable management strategy is to slow the rate of spread. However, to be effective, this will require detailed spread data and an understanding of the influence of environmental conditions and landscape structure on invasion rates. We used a time series of remotely sensed distribution maps and a spatial simulation model to study spread of the invasive Lepidium latifolium (perennial pepperweed) in California's Sacramento-San Joaquin River Delta. L. latifolium is a noxious weed and exhibited rapid, explosive spread. Annual infested area and empirical dispersal kernels were derived from the remotely sensed distributions in order to assess the influence of weather conditions on spread and to parameterize the simulation model. Spread rates and dispersal distances were highest for nascent infestations and in years with wet springs. Simulations revealed that spread rates were more strongly influenced by the length of long-distance dispersal than by temporal variation in its likelihood. It is thus important to capture long-distance dispersal and the conditions that facilitate spread when collecting data to parameterize spread models. Additionally, management actions performed in high-spread years, targeting long-distance recruits, can effectively contain infestations. Corridors were relatively unimportant to spread rates; their effectiveness at enhancing rate of spread was limited by the species' dispersal ability and the time needed to travel through the corridor. In contrast, habitat abundance and shape surrounding the introduction site strongly influenced invasion dynamics. Satellite patches invading large areas of invasible habitat present especially high risk.

Heterogeneity in the distribution of genetically modified and conventional oilseed rape within fields and seed lots

The implementation of co-existence in the commercialisation of GM crops requires GM and non-GM products to be segregated in production and supply. However, maintaining segregation in oilseed rape will be made difficult by the highly persistent nature of this species. An understanding of its population dynamics is needed to predict persistence and develop potential strategies for control, while to ensure segregation is being achieved, the production of GM oilseed rape must be accompanied by the monitoring of GM levels in crop or seed populations. Heterogeneity in the spatial distribution of oilseed rape has the potential to affect both control and monitoring and, although a universal phenomenon in arable weeds and harvested seed lots, spatial heterogeneity in oilseed rape populations remains to be demonstrated and quantified. Here we investigate the distribution of crop and volunteer populations in a commercial field before and during the cultivation of the first conventional oilseed rape (winter) crop since the cultivation of a GM glufosinate-tolerant oilseed rape crop (spring) three years previously. GM presence was detected by ELISA for the PAT protein in each of three morphologically distinguishable phenotypes: autumn germinating crop-type plants (3% GM), autumn-germinating 'regrowths' (72% GM) and spring germinating 'small-type' plants (17% GM). Statistical models (Poisson log-normal and binomial logit-normal) were used to describe the spatial distribution of these populations at multiple spatial scales in the field and of GM presence in the harvested seed lot. Heterogeneity was a consistent feature in the distribution of GM and conventional oilseed rape. Large trends across the field (50 x 400 m) and seed lot (4 x 1.5 x 1.5 m) were observed in addition to small-scale heterogeneity, less than 20 m in the field and 20 cm in the seed lot. The heterogeneity was greater for the 'regrowth' and 'small' phenotypes, which were likely to be volunteers and included most of the GM plants detected, than for the largely non-GM 'crop' phenotype. The implications of the volunteer heterogeneity for field management and GM-sampling are discussed.