Sampling for weed spatial distribution mapping need not be adaptive

Weeds are species of interest for ecologists because they are competitors of the crop for resources but they also play an important role in maintaining biodiversity in agroecosystems. To study their spatial distribution at the field scale, only sampled observations are available due to the cost of sampling. Weeds sampling strategies are static. However, in the domain of spatial sampling, adaptive strategies have also been developed with, for some of them, an important on-line or off-line computational cost. In this article we provide answers to the following question: Are the current adaptive sampling methods efficient enough to motivate a wider use in practice when sampling a weed species at a field scale? We provide a comparison of the behaviour of 8 static strategies and 3 adaptive ones on four criteria: density class estimation, map restoration, spatial aggregation estimation, and sampling duration. From two weeds data sets, we estimated six contrasted Markov Random Field (MRF) models of weed density class spatial distribution and a model for sampling duration. The MRF models were then used to compare the strategies on a large set of simulated maps. Our main finding was that there is no clear gain in using adaptive sampling strategies rather than static ones for the three first criteria, and adaptive strategies were associated to longer sampling duration. This conclusion points out that for weed mapping, it is more important to build a good model of spatial distribution, than to propose complex adaptive sampling strategies.

[1]  L. K. Ward,et al.  The role of weeds in supporting biological diversity within crop fields , 2003 .

[2]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Mathieu Bonneau,et al.  Reinforcement learning-based design of sampling policies under cost constraints in Markov random fields: Application to weed map reconstruction , 2014, Comput. Stat. Data Anal..

[4]  Alfred Stein,et al.  Analyzing spatial count data, with an application to weed counts , 2007, Environmental and Ecological Statistics.

[5]  D. J. Brus,et al.  Sampling for Natural Resource Monitoring , 2006 .

[6]  E. Oerke.  Crop losses to pests , 2005, The Journal of Agricultural Science.

[7]  Sabrina Gaba,et al.  Non-random distribution of weed species abundance in arable fields , 2012 .

[8]  Anders Brix,et al.  Space‐time Multi Type Log Gaussian Cox Processes with a View to Modelling Weeds , 2001 .

[9]  J. Besag.  Spatial Interaction and the Statistical Analysis of Lattice Systems , 1974 .

[10]  Regis Chikowo,et al.  Integrated Weed Management systems allow reduced reliance on herbicides and long-term weed control , 2009 .

[11]  Hong S. He,et al.  An aggregation index (AI) to quantify spatial patterns of landscapes , 2000, Landscape Ecology.

[12]  Giuseppe Zanin,et al.  Incorporation of weed spatial variability into the weed control decision‐making process , 1998 .

[13]  C. Ji,et al.  A consistent model selection procedure for Markov random fields based on penalized pseudolikelihood , 1996 .

[14]  Lisa Norton,et al.  Environmental and management factors determining weed species composition and diversity in France , 2008 .

[15]  Jonathan Storkey,et al.  Using Assembly Theory to Explain Changes in a Weed Flora in Response to Agricultural Intensification , 2010, Weed Science.

[16]  Daniel Spring,et al.  Model-based adaptive spatial sampling for occurrence map construction , 2013, Stat. Comput..

[17]  Colbach,et al.  Evaluating field-scale sampling methods for the estimation of mean plant densities of weeds , 2000 .

[18]  J. J. de Gruijter,et al.  A hybrid design-based and model-based sampling approach to estimate the temporal trend of spatial means , 2012 .

[19]  J. Cardina,et al.  The nature and consequence of weed spatial distribution , 1997, Weed Science.

[20]  J. Cardina,et al.  Analysis of Spatial Distribution of Common Lambsquarters (Chenopodium album) in No-Till Soybean (Glycine max) , 1995, Weed Science.

[21]  Sabrina Gaba,et al.  Trajectories of weed communities explained by traits associated with species’ response to management practices , 2012 .

[22]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[23]  L. Wiles,et al.  Spatial Distribution of Broadleaf Weeds in North Carolina Soybean (Glycine max) Fields , 1992, Weed Science.

[24]  C. Nicholls,et al.  Plant biodiversity enhances bees and other insect pollinators in agroecosystems. A review , 2013, Agronomy for Sustainable Development.

[25]  Donald Geman,et al.  Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images , 1984 .

[26]  Brett Whelan,et al.  Sampling strategy is important for producing weed maps: a case study using kriging , 2002, Weed Science.

[27]  David E. Clay,et al.  Spatial distribution, temporal stability, and yield loss estimates for annual grasses and common ragweed (Ambrosia artimisiifolia) in a corn/soybean production field over nine years , 2006, Weed Science.

[28]  Vincent Bretagnolle,et al.  Weeds for bees? A review , 2015, Agronomy for Sustainable Development.

[29]  Francisca López Granados.  Weed detection for site-specific weed management: Mapping and real-time approaches , 2011 .