40% of the territory of Western Siberia is occupied by solonetzes, used mainly for growing forage grasses. The phytomeliorative effect of yellow sweet clover (cultivated in crop rotation at the Siberian Federal Research Center of Agrobiotechnologies (SFSCA) of the Russian Academy of Sciences (RAS) for 31 years), as well as post crop rotation grassing with a mixture of brome and alfalfa (the mixture was sown after twenty years of crop rotation) on the microflora of medium solonetz was studied. It is shown that the cultivation of phytomeliorants has led to an increase in the representation of classes and orders of bacteria in the microbiome of the medium solonetz, associated with nitrogen fixation and mineralization activity of the soil in relation to nitrogen-containing substances. In the post crop rotation grassing variant, a low bacterization of the soil with Verrucomicrobia and a high, as in virgin soil, Acidobacteria were found.

Traditional weed management approaches can become more effective when integrated with artificial intelligence (AI) models. The identification and classification of weeds using AI techniques can play an important role in weed control, helping to increase crop yields. The rapid development of deep learning methods based on convolutional neural networks helps in solving this problem. In particular, the trained algorithm is able to automatically extract information from images, detect and classify weeds. The article discusses the construction of an image classifier using the ResNet-18 architecture. Three types of weeds are present in buckwheat crops with different intensity: wild oat (Avena fatua), field bindweed (Convolvulus arvensis) and barnyard grass (Echinochloa crus-galli). The task of the classifier is to recognize these weeds in the photograph and determine one of the two gradations of weediness of the site—the number of weeds exceeds the economic injury level (EIL) or does not exceed. The effectiveness of the proposed algorithm is confirmed by the rather high quality of the classifier predictions (the number of correct classifications for the initial set of 24 images is 87%) and the construction of the Confusion matrix.

In the Baraba Plain, the influence of crop rotations with sweet clover and awnless bromegrass on solonets has been studied in dynamics for more than 30 years. It has been noted that in the soil with phytomeliorative crop rotations total salt reserves have significantly decreased in comparison with the initial virgin soil. In the 0–20 cm soil layer, their number decreased 3.8–4.4 times, in the 20–40 cm layer – 4.6–7.7 times. As a result of grassing of the phytomeliorative crop rotation plots with a mixture of awnless bromegrass and alfalfa blue-hybrid, the effect of desalinization is decreasing (in the upper layer on average by 6.4 and 9.3%, in the lower layer – by 24.9% in the aftermath of the crop rotation with awnless bromegrass). The identified changes in the soil salinity have been reflected in the representation of salt-tolerant and salt-sensitive bacteria. The abundance of low salt-tolerant representatives of the class Spartobacteria on the grassed area after crop rotations with sweet clover and bromegrass decreased by 3.2 and 3.6 times, and the abundance of the relatively salt-loving Cytophagia increased by 1.6 and 2.4 times. In the sown meadow after crop rotation with sweet clover, a higher amount of complexly decomposable plant residues (mainly cereals) was observed, as evidenced by the increased content of acidobacteria. According to the abundance of the genera Gaiella from the class Thermoleophilia and Microlunatus from the class Actinobacteria, the meliorative effect in terms of desalinization and aeration of solonets is greater in sweet clover than in bromegrass. Grassing increases mineralization activity and oligotrophic soil in solonetz on average in 20–40 cm layer more strongly than in 0–20 cm layer by 1.6–2.2 times. Potential microbiological humus accumulation under sown meadow decreases in the upper layer of the plot previously occupied by the rotation with sweet clover, and in the lower layer – by the rotation with bromegrass.

Vast areas of Western Siberia (about 40%) are occupied by saline soils. Solonetzic agricultural lands are mainly used for growing perennial grasses - phytomeliorants. To increase the yield of fodder crops on solonetz lands, specialists of the Siberian Research Institute of Fodder Crops in the 80s of the last century developed phytomeliorative crop rotations. The article considers the effect of grassing phytomeliorative crop rotations with a mixture of brome (Bromus inermis Leyss.) and alfalfa (Medicago varia Mart.) on the microflora of meadow solonetzes (hydromorphic) (Gleyic Solonetz Albic) (for 13 years). It is shown that post-rotational grassing led to a significant desalinization of the upper soil horizon and a decrease in its alkalinity, as well as to an improvement in the water-air regime of solonetzes. Long-term cultivation of fodder crop rotations and subsequent grassing formed a specific soil microbiome, characterized by taxonomic diversity of microorganisms and a greater proportion of copiotrophs in the dominant phyla, which indirectly indicates an increase in the carbon content and nitrogen available to plants in the phytomeliorated solonetz.

Traditional methods of monitoring weeds in crops are not suitable for integration with modern “smart” agricultural machinery. Automating the process of accurately identifying the species composition of weeds in each field will be an important step for the development of a plant protection system, contributing to higher yields. Deep learning (DL) models, which have become widespread in recent years, are successfully helping to solve this complex agricultural problem. In this study, we built a classifier based on the ResNet-18 deep learning model, which is able to detect weeds with the corresponding weediness gradations in photographs from field plots with oil flax (Linum usitatissimum L.). There are 4 types of weeds in crops of oil flax with different intensity—field bindweed (Convolvulus arvensis), white goosefoot (Chenopodium album), leafy spurge (Euphorbia virgata), and wild buckwheat (Fallopia convolvulus). The task of the classifier is to recognize these weeds in the photograph and determine one of the two gradations of weediness of the plot - the number of weeds exceeds the economic injury level (EIL) or does not exceed. The models were trained with different epoch values (10, 20, 30), the accuracy of which ranged from 72.5 to 93.3%.

In the modern era of urbanization, air pollution is becoming a significant cause of changes in the health of the population. The contribution of the complex of atmospheric pollutants to the state of the urban environment can be assessed by the bioindication method. In the article, this is done using the fluctuating asymmetry of the leaf blades of the silver birch on the example of the one and a half million city of Novosibirsk and a rural settlement located 50 km away. Novosibirsk is an industrial center and has a high transport and logistics significance. Bioindication data testify to the critical state of its environment in places of high transport load and industrial congestion and to a significant deviation from the norm when moving away from the park zone. This requires the development of additional measures aimed at minimizing the impact of environmental risks on the population.

В данной работе представлены результаты разработки имитационной модели уборочных работ сельскохозяйственных культур в среде AnyLogic. Работа выполнена на данных конкретного землепользования по внутрихозяйственной логистике одного из предприятий Новосибирской области (ОС Элитная, р.п. Краснообск: 54°54'57"с.ш., 82°57'6"в.д.). В основу модели заложены математические зависимости сроков созревания основных сельскохозяйственных культур в зависимости от их требований к условиям теплообеспеченности территории (сумм температур воздуха ∑t>10ºC), а также выполнена кластеризация культур по их назначению (кормовые и зерновые). В основе расчётного модуля также использовали параметры машинно-технологического обеспечения предприятия: ширина жатки комбайнов, объем бункера комбайнов и кузовов обслуживающих транспортных средств (ОТС), площади участков, назначение и урожайность сельскохозяйственных культур. Описано построение структуры работы взаимодействия комбайнов с ОТС, учитывая их технические характеристики, фактические и прогнозные данные суточных метеопараметров. Продемонстрирована возможность использования имитационной среды AnyLogic для решения частных задач в области сельскохозяйственного производства. Имитационная модель может быть использована в качестве инструмента поддержки принятия решений при планировании тактики и стратегии проведения уборочных работ, а также оптимизации уборочно-транспортных процессов в растениеводстве. This paper presents the results of the development of a simulation model of harvesting crops in the AnyLogic environment. The work was carried out on the data of a specific land use for on-farm logistics of one of the enterprises of the Novosibirsk region (Elite OS, Krasnoobsk: 54°54'57" s.w., 82°57'6" v.d.). The model is based on mathematical dependences of the maturation dates of the main crops depending on their requirements for the conditions of heat supply of the territory (sums of air temperatures ∑t>10ºC), and clustering of crops according to their purpose (fodder and grain) is also performed. The calculation module was also based on the parameters of the machine and technological support of the enterprise: the width of the harvester harvester, the volume of the hopper of the harvesters and the bodies of the servicing vehicles (OTS), the area of the plots, the purpose and yield of agricultural crops. The construction of the structure of the interaction of combines with OTS is described, taking into account their technical characteristics, actual and forecast data of daily meteorological parameters. The possibility of using the AnyLogic simulation environment for solving particular problems in the field of agricultural production is demonstrated. The simulation model can be used as a decision support tool when planning tactics and strategies for harvesting operations, as well as optimizing harvesting and transport processes in crop production.

Авторами рассматривается возможность использования нейросетевой модели (FFNN – нейронная сеть прямого распространения), для прогнозирования урожайности яровой пшеницы в условиях лесостепи Западной Сибири. В исследовании использованы материалы длительных полевых опытов СибНИИЗиХ – структурного подразделения СФНЦА РАН, проведенные в северной лесостепи Приобья, а также данные о метеорологических показателях Новосибирского поста метеонаблюдений за 2001-2018 гг. Работа выполнена с использованием общедоступных данных для универсальности системы при ее использовании в различных природно-сельскохозяйственных условиях. В качестве предикторов выделены качественные факторы (система обработки почвы, предшествующая культура, размещение культуры после пара применение средств интенсификации) и метеорологические показатели (среднедекадные температуры воздуха и суммы осадков), определяющие урожайность культуры на исследуемой территории. Выполнено построение модели, позволяющей осуществить прогноз урожайности яровой пшеницы на будущий вегетационный период в зависимости от заданных параметров. Коэффициент детерминации модели составил 0.93, а средняя абсолютная ошибка изменялась в пределах 0.05±0.03, что являются достаточно высоким результатом точности предиктивных моделей в постоянно изменяющихся условиях при совокупности абиотических факторов и управляющего воздействия. Полученные в ходе работы теоретические и практические результаты могут быть использованы при разработке систем поддержки принятия решений, а также при планировании и оценке эффективности размещения сельскохозяйственного производства растениеводческой продукции в изменяющихся погодно-климатических условиях на территории лесостепи Приобья. The authors consider the possibility of using a neural network model (FFNN – feed forward neural network) to predict the yield of spring wheat in the forest-steppe of Western Siberia. The study involved materials from long–term field experiments of SibNIIZiH, a structural subdivision of the SFSCA RAS, conducted in the northern forest-steppe of the Ob region, as well as data on meteorological indicators of the Novosibirsk meteorological observation post for 2001-2018. The work was carried out using publicly available data for the universality of the system when it is used in various natural and agricultural conditions. Qualitative factors (the tillage system, the previous crop, the placement of the crop after steam, the use of intensification means) and meteorological indicators (average decadal air temperatures and precipitation amounts) that determine the crop yield in the study area are identified as predictors. A model has been constructed that allows forecasting the yield of spring wheat for the future growing season, depending on the specified parameters. The coefficient of determination of the model was 0.93, and the mean absolute error varied within 0.05±0.03, which is a fairly high result of the accuracy of predictive models in constantly changing conditions with a combination of abiotic factors and control action. The theoretical and practical results obtained in the course of the work can be used in the development of decision support systems, as well as in planning and evaluating the effectiveness of the placement of agricultural production of crop production in changing weather and climatic conditions on the territory of the Ob region.

This paper presents the possibility and feasibility of using the Bayesian network method and multinomial logistic regression to predict the yield of spring wheat. To build and train the model the data of a long-term multifactorial stationary field experiment of the Siberian Research Institute of Husbandry and Chemicalization of Agriculture (SRIHCA) of the Siberian Federal Scientific Centre of Agro-BioTechnologies (SFSCA) of the Russian Academy of Sciences (RAS) for the time period of 2004–2018 were used. During the analysis of the data sample, the main predictors of the model affecting the spring wheat yield were identified. The predictors are represented by qualitative and quantitative parameters of the working area: predecessor, tillage, HTC (weather conditions), pesticides, and yield by appropriate gradations (events). As a result, models were built and tested that were able to predict the yield of spring wheat, depending on the prevailing conditions. To assess the predictive ability of the models were tested on the original sample. In Bayesian Networks, the total share of correct forecasts for all categories of spring wheat yields is 81%. The total share of correct forecasts obtained by implementing the multinomial regression model is 83%. The constructed models make it possible to predict with acceptable reliability.

The use of intelligent technologies for the systematization of agrotechnological knowledge, by creating a specialized software system for storing and presenting knowledge in the field of crop production, represents one of the most promising areas for improving the efficiency of agricultural production planning. The paper deals with the relevant issues of creating a knowledge base for information support system of the operator’s activities on the mineral fertilizer application. The increased requirements for modern agriculture to be resource-saving and environmentally friendly stem from the development of this direction. The software LogicGem, an easy-to-use tool for creating, editing, checking and compiling decision table logic, was used to build a knowledge base.

The article presents the results of a five-year study of plant-microbe interactions in a small medium sodium solonetz. The research was carried out in the Barabinsk lowland of Siberia, where Bromus inermis was cultivated for 33 years in the fodder crop rotation. Its phytomelioration effect was reflected in soil desalinization which was proved by decrease of EC in 0-20 cm layer by 11.3 times, decrease of its oligotrophy by 6.3 times, replenishment of solonetz nitrogen fund and intensification of microbiological processes of humus accumulation. The taxonomic structure of microbiome under the crop rotation impact with B. inermis altered towards increasing the representation of genera from alpha-, beta- and delta-proteobacteria classes, functionally associated with the improvement of soil fertility, Gemmatimonadetes as a dryness indicator and reducing the Acidobacteria groups Gp4 and Gp6, associated with the pH control.

Today, the effective and sustainable operation of agricultural production requires the deployment of information systems using a set of digital technologies. Predictive analytics in these systems shall take the dominant position since it is difficult or almost impossible to make a correct decision on its management without forecasting the transformation of conditions, objects and processes occurring in agriculture. The paper deals with a decision support system for use by specialists in crop production. The guidelines of farm management on the basis of key nodes of production decision-making are highlighted over the entire period of agronomic work. The structure includes 12 key nodes, 28 models, 1 database and 415 agricultural indicators available to the user. The system issues a document with a recommendation to the producer for each node. On the basis of forecast models of agrometeorological resource and spring wheat yield, planned models are built. They allow calculating the time, analyzing the economic performance and interaction of machinery in various agro-mechanical activities throughout the agricultural season. Databases of crop varieties (hybrids), pesticides and fertilizers relevant to the territory of the Russian Federation have been developed. Models and methods were tested on one of experimental farms of Novosibirsk region.

This paper presents multinomial logistic regression and neural network models that allow predicting and thereby quickly determining the content of nitrate nitrogen in the 0-40 cm soil layer before sowing. To train the models, we used the data of a long-term multifactorial field experiment of the Siberian Research Institute of Agriculture and Chemistry of the Siberian Federal Scientific Research Center of the Russian Academy of Sciences for 2009-2018, established in 1981 on leached chernozem in the OS "Elitnaya" - a branch of the Siberian Federal Scientific Research Center of the Russian Academy of Sciences. Taking into account the features of the statistical sample (observation data and analyses), the main predictors of the models that affect the content of nitrate nitrogen in the soil (target indicator) are determined, they are presented qualitative (predecessor, tillage) and quantitative (weather conditions and the content of productive moisture before sowing in the layer 0-100 cm) by factors with appropriate gradations. The models showed a fairly high reliability when verified on empirical data and can be used as a tool for forecasting. The quality of the developed multinomial logistic regression model was assessed using the coefficient of determination, which was 78% according to the Nagelkerke measure, and 72% according to the Coxsei Snell measure. To determine the predictive ability of the neural network, an ROC analysis was carried out, which showed that the area under the ROC curve for each category of the target indicator was close to 1, which indicates a high predictive power of this method. A comparative assessment of the predictive capabilities of the trained models was carried out. The overall proportion of correct predictions for multinomial logistic regression is 80.6%, in the neural network model 89.5%. Keywords: MULTINOMIAL LOGISTIC REGRESSION, NEURAL NETWORK, NITRATE NITROGEN, SOIL

The possibilities and feasibility of using the Bayesian network of trust and logistic regression to predict the content of nitrate nitrogen in the 0-40 cm soil layer before sowing have been investigated. Data from long-term multifactor field experience at the Siberian Research Institute of Farming and Agricultural Chemization of SFSCA RAS for 2013-2018 were used to train the models. The experiment was established on leached chernozem in the central forest-steppe subzone in 1981 in the Novosibirsk region. Considering the characteristics of the statistical sample (observation and analysis data), the main predictors of the models affecting nitrate nitrogen content in soil were identified. The Bayesian trust network is constructed as an acyclic graph, in which the main (basic) nodes and their relationships are denoted. Network nodes are represented by qualitative and quantitative plot parameters (soil subtype, forecrop, tillage, weather conditions) with corresponding gradations (events). The network assigns a posteriori probability of events for the target node (nitrate-nitrogen content in the 0-40 cm soil layer) as a result of experts completing the conditional probability table, taking into account the analysis of empirical data. Two scenarios were analyzed to test the sustainability of the network and satisfactory results were obtained. The result of the logistic regression is the coefficients characterizing the closeness of the relationship between the dependent variable and the predictors. The coefficient of determination of the logistic regression is 0.7. This indicates that the quality of the model can be considered acceptable for forecasting. A comparative assessment of the predictive capabilities of the trained models is given. The overall proportion of correct predictions for the Bayesian confidence network is 84%, for logistic regression it is 87%.