Search

Crop yield forecasting is critical in modern agriculture to ensure food security, economic stability, and effective resource management. The main goal of this study was to combine historical multisource satellite and environmental datasets with a deep learning (DL) model for soybean yield forecasting in the United States' Corn Belt. The following Moderate Resolution Imaging Spectroradiometer (MODIS) products were aggregated at the county level. The crop data layer (CDL) in Google Earth Engine (GEE) was used to mask the data so that only soybean pixels were selected. Several machine learning (ML) models were trained by using 5 years of data from 2012 to 2016: random forest (RF), least absolute shrinkable and selection operator (LASSO) regression, extreme gradient boosting (XGBoost), and decision tree regression (DTR) as well as DL-based one-dimensional convolutional neural network (1D-CNN). The best model was determined by comparing their performances at forecasting the soybean yield in 2017-2021 at the county scale. The RF model outperformed all other ML models with the lowest RMSE of 0.342 t/ha, followed by XGBoost (0.373 t/ha), DTR (0.437 t/ha), and LASSO (0.452 t/ha) regression. However, the 1D-CNN model showed the highest forecasting accuracy for the 2018 growing season with RMSE of 0.280 t/ha. The developed 1D-CNN model has great potential for crop yield forecasting because it effectively captures temporal dependencies and extracts meaningful input features from sequential data. [ABSTRACT FROM AUTHOR]

Wildfire poses a serious risk in several regions of the world threatening urban, agricultural areas and natural ecosystems as well. Nature conservation has important role to be prepared for the management of postfire environmental degradation and restoration for protected areas preserving valuable ecosystems. The improving temporal and spatial resolution of remote sensing and GIS methods significantly contributes to map the changes for accelerating management steps of restoration. In this study a severe wildfire and its impacts were assessed in case of a protected area of the Kiskunság National Park in Hungary, which was partly burnt down in 2012. The aim of this research was to efficiently and accurately assess the damages and to plan and execute the restoration works using remote sensing tools. Aerial data collection was performed one month, and one year after the fire. In 2014 the regenerated vegetation was surveyed and mapped in the field. Using the aerial photographs and the field data, the degree and extent of the fire damages, the types and the state of the vegetation and the presence and proportion of the invasive species were determined. Semi-automatic methods were used for the classification of completely, partially damaged and undamaged areas. Based on the results, the reforestation of the burnt area is suggested to prevent the overspreading of white poplar against common junipers and to clean the area from the most frequent invasive species. To monitor the regeneration of the vegetation and the spreading of the invasive species, further aerial photography and field campaigns are planned.

25% and 40% of territory of Hungary is moderate to highly vulnerable to deflation. However, precise estimates about the soil loss and related losses of organic matter and nutrients due to wind erosion are missing in most cases. In order to determine magnitudes of nutrient masses removed at wind velocities that frequently occur in SE Hungary, in-situ experiments using a portable wind tunnel have been conducted on small test plots with an erosional length of 5.6 m and a width of 0.65 m. The wind tunnel experiments have been carried through on a Chernozem which is typical for this region. In order to compare the effects of soil coverage on the masses of blown soil sediment and adsorbed nutrients, two soil surface types have been tested under similar soil moisture und atmospheric conditions: (1) bare soil (dead fallow) and (2) bare soil surface interrupted by a row of maize plants directed downwind along the center line of the test plots. The results of our experiments clearly show that a constant wind velocity of 15 m s-1 (at a height of 0.3 m) lasting over a short time period of 10 minutes can already cause noticeable changes in the composition and size of soil aggregates at the top of the soil surface. Due to the grain size selectivity of the erosive forces the relative share of soil aggregates comprising diameters > 1 mm increased by 5-10% compared with the unaffected soil. Moreover it has shown that short time wind erosion events as simulated in this study can result in erosion rates between 100 and 120 g m-2, where the erosion rates measured for bare soils are only slightly, but not significantly higher than those of the loosely vegetated ones. Soil samples taken from sediment traps mounted in different heights close to the outlet of the wind tunnel point to an enrichment of organic matter (OM) of about 0.6 to 1 % by mass referred to the control samples. From these findings has been calculated that the relocation of organic matter within short term wind erosion events can amount to 4.5 to 5.0 g OM m-2. With the help of portable field wind tunnel experiments we can conclude that our valuable, high quality chernozems are struck by wind erosion mainly in drought periods.

Inland excess water floodings are a common problem in the Carpathian Basin. Nearly every year large areas are covered by water due to lack of natural runoff of superfluous water. To study the development of this phenomenon it is necessary to determine where these inundations are occurring. This research evaluates different methods to classify inland excess water occurrences on a study area covering south-east Hungary and northern Serbia. The region is susceptible to this type of flooding due to its geographical circumstances. Three separate methods are used to determine their applicability to the problem. The methods use the same input data set but differ in approach and complexity. The input data set consists of a mosaic of RapidEye medium resolution satellite images. The results of the classifications show that all three methods can be applied to the problem and provide high quality satellite based inland excess water maps over a large area.

Copyright of Agrochemistry & Soil Science / Agrokémia és Talajtan is the property of Akademiai Kiado and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)