Abstract: Statistical learning is a fairly new term referring to a set of supervised and unsupervised modelling and prediction techniques. It is based on traditional statistics but has been highly enhanced inspired by developments in machine learning and data mining. The main focus of statistical learning is to estimate the functions that quantify relations between several parameters and observed responses. These functions are further used for prediction, inference or a combination of both. For a particular case of quantitative responses, regularization techniques in regression are developed to overcome the weaknesses of ordinary least square (OLS) regression in prediction. These new shrinkage methods outperform OLS for prediction, especially in large datasets. In this study, a large dataset of field experiments on winter oilseed rape in Denmark for 22 years (1992 to 2013) was collected. Biweekly climatic data along with sowing date, harvest date, soil type and previous crop are considered as the explanatory variables. Yield of winter oilseed rape is considered as response variable. LASSO and Elastic Nets are the regularization techniques used to estimate the functions. Hold-one-out cross validation method for testing the prediction power reveals that these techniques are much useful in both prediction and inference. Since these techniques are included in recent versions of some software packages (e.g. R), they can be easily implemented by users at any level. The estimated function (model) is further used to predict the oilseed rape yield responses to climate change for several scenarios using representative weather data produced by a weather generator.

Abstract: The assessment of net ecosystem exchange and respiration of ecosystem of terrestrial ecosystems is necessary to improve our knowledge about carbon cycle in nature. Here we present measurements of CO2 fluxes for a winter wheat temperate climate ecosystem (buckwheat in the previous years) located in the Lubelskie province (eastern Poland) using a closed dynamic chamber system over a 2013 vegetation season. Measurements of carbon dioxide emission from soils and its assimilation by plants were carried out on a typical for Lubelskie highland arable land located in the Stany Nowe (N50o49’17.0555”, E22o16’28.51”, height 243m above sea level) using the set of two chambers (transparent and dark). Carbon dioxide fluxes have been measured by EGM-4 PP Systems sensor during fixed stages of the plant growing season. During the experiment carbon emission from soil ranged from 151 to 764 mg C·m-2·h-1 and its assimilation by plants ranged from -148 (emission) to 1585 mg C·m-2·h-1. We found substantial differences in emission and assimilation of carbon in the winter wheat ecosystem. This, along with other measurements (meteorological factors and soil and plant parameters) carried out in the Stany Nowe can be used as a high quality data to verify various models of  emission of greenhouse gases. The chamber technique occurs to be  a useful tool for determining carbon dioxide exchange between ecosystem surface and the atmosphere.

Abstract: Interactions of climate, soil and management practices in cropping systems can be simulated at different scales to provide information for decision making. Low resolution simulation need less effort, but important details could be lost through data aggregation effects (DAEs). This paper aims to provide a general method to assess the DAEs on weather data and the simulation of cropping systems, and further investigate how the DAEs vary with changing crop models, crops, variables and production systems. A 30-year continuous cropping system was simulated for winter wheat and silage maize and potential, water-limited and water-nitrogen-limited production situations. Climate data of 1 km resolution and aggregations to resolutions of 10 to 100 km was used as input for the simulations. The data aggregation narrowed the variation of weather data and DAEs increased with increasingly coarser spatial resolution, causing the loss of hot spots in simulated results. Spatial patterns were similar across different resolutions. Consistent with DAEs on weather data, the DAEs on simulated yield (0 to 1.2 t ha-1 for winter wheat and 0 to 1.7 t ha-1 for silage maize), evapotranspiration (3 to 45 mm yr-1 for winter wheat and 4 to 40 mm yr-1 for silage maize), and water use efficiency (0.02 to 0.25 kg m-3­ for winter wheat and 0.04 to 0.4 kg m-3­ for silage maize), increased with coarser spatial resolution. Thus, if spatial information is needed for local management decisions, higher resolution is needed to adequately capture the spatial heterogeneity or hot spots in the region.