Wet soils play an important role in hydrological, biological and chemical processes, and knowledge on their spatial distribution is essential in forestry, agriculture and similar fields. Digital elevation models (DEM) and various hydrological indexes are used to perform water runoff and accumulation processes. The prerequisite for the calculation of the hydrological indexes is the most accurate representation of the Earth’s surface in the DEM, which must be corrected as necessary to remove surface artefacts that create a dam effect. In addition, different resolutions for DEM give different results, so it is necessary to evaluate what resolution data is needed for a particular study. The aim of this study is to evaluate the feasibility of using existing ditch vector data for DEM correction and the resulting implications for soil moisture prediction. Applied methodology uses a network of available ditch vectors and creates gaps in the overlapping parts of the DEM. The data were processed using open source GIS software QGIS, GRASS GIS and Whitebox GAT. Ditch vector data were obtained from JSC Latvian State Forests and the Latvian Geospatial Information Agency. The results show that by applying the bottomless ditch approach in forest lands on moraine deposits, depending on the accuracy of the ditch vector data, the values of the prediction of the soil wetness both increase and decrease. On the other hand, in forest lands on graciolimnic sediments it is visible that predicted soil wetness values increase in the close proximity of ditches. For forest lands on glaciofluvial and eolitic sediments there were no visible changes because of lack of ditches.

Precision of the forest inventory still is one of the most important problems in the forestry nowadays. The aim of this research was to estimate the results of the combined forest inventory (CFI), using high spatial resolution aerial images in the planned areas of clear-cuts, comparing the results with the calipering and production files of harvesters. Testing of algorithms showed considerable difference in results between the CFI, forest inventory data and harvester production data. CFI results and production data had a close correlation with R2 =0.83. Comparing CFI calculated growing stock with production data, the average relative error amounted to 10.7%, which means the possibility for integration of these results into the forest inventory system. Comparing to CFI, there is a weak correlation between forest inventory and production data with R2 =0.34. The results indicate that LiDAR CFI technology can be used in the forecasting of the forest management, offering precise information about potential amount and economic value of assortments.

Nowadays, the use of Unmanned Aerial Vehicle flying at a low altitude in conjunction with photogrammetric and LiDAR technologies allows to collect images of very high-resolution to generate dense points cloud and to simulate geospatial data of territories. The technology used in experimental research contains reconstruction of topography of surface with historical structure, observing the recreational infrastructure, obtaining geographic information for users who are involved in preservation and inspection of such unique cultural/ heritage object as are mounds in Lithuania. In order to get reliable aerial mapping products of preserved unique heritage object, such photogrammetric/ GIS procedures were performed: UAV flight for taking images with the camera; scanning surface by LiDAR simultaneously; processing of image data, 3D modelling and generation of orthophoto. Evaluation of images processing results shows that the accuracy of surface modelling by the use of UAV photogrammetry method satisfied requirements – mean RMSE equal to 0.031 m. The scanning surface by LiDAR from low altitude is advisable, relief representation of experimental area was obtained with mean accuracy up to 0.050 m. Aerial mapping by the use of UAV requires to specify appropriate ground sample distance (GSD) that is important for reducing number of images and time duration for modelling of area. Experiment shows that specified GSD of 1.7 cm is not reasonable; GSD size increased by 1.5 times would be applicable. The use of different software in addition for DSM visualization and analysis is redundant action.

Wet areas in agricultural lands are usually not fully or properly managed due to problematic accessibility by heavy machinery and are associated with lower crop yields. There are neither studies regarding spatial distribution of wet agricultural areas in Latvia nor large scale soil maps. Being aware of these wet areas, it would be possible to plan actions for effective management of these areas, starting with a scale of landscape. A geographic information system model could serve as an assistant for decision-making, such as, a direct support for the management of amelioration systems, change of land use and management patterns or granting support payments. Remote sensing data like Sentinel-2 satellite images and LiDAR (Light detecting and ranging) technology can be used to identify local wet areas. The focus of this article is to evaluate different remote sensing indices and methods that can be used to identify wet areas in agricultural lands using open access data and software. From 52 indices, which were analysed with soil moisture field measurements in 33 sample plots, only two of them showed statistical significance in linear regression model (p is less than 0.05): normalized height model in resolution of 25 meters (r2 =0.45) and visible blue spectral band in April (r2 =0.39). Results from this study help to focus on different aspects of remote sensing data usage and methodology for future improvements in order to fully implement LiDAR and Sentinel-2 data for identification of wet areas in agricultural lands.

Road landscapes can be considered important resources for place development. They create impression about the infrastructure of places and transport, which is an important aspect of attracting investment and tourism development. Yet this field of landscape planning and design is hardly studied and needs more attention in Latvia. Institutions at different planning levels and from several fields of expertise are involved in road landscape development. In order to achieve successful cooperation among all the parties involved, it is necessary to reflect the information about road landscape development in the way that it can be easily perceived and understood. Studies in landscape perception prove that people perceive visual information about landscape design and planning better than textual information and regular maps. The purpose of the paper is to introduce with a method of three dimensional (3D) road landscape modelling, developed by authors as a tool for road landscape design aesthetic evaluation, which can be used to demonstrate design variants to wider public and stakeholders. We demonstrate what kind of data are necessary for road landscape modelling, how they are obtained and processed, why certain modelling programs are chosen. The methodology, problems, which occurred during the modelling, and the chosen solutions are described. Results show that chosen methodology is appropriate for large scale projects. The experience gained from the project helps to evaluate the suitability of certain computer programs for road landscape planning and design.

This study investigates the distribution and carbon content of organic soils in Latvia, leveraging machine learning techniques alongside remote sensing and National Forest Inventory (NFI) data to enhance the precision of organic soil mapping. Our approach integrates data from various sources, including airborne laser scanning (ALS) data, digital elevation models (DEM), depth-to-water (DTW) and wet area maps (WAM), and historical organic soil data. By classifying over 24,000 soil probing measurements across Latvia into distinct peat layer thickness categories, we develop a machine learning model that categorizes the thickness of the organic layer with notable accuracy. Our findings indicate that the model, particularly when employing the xgbTREE algorithm and over-sampling method, successfully identifies areas with peat layers thicker than 40 cm, demonstrating a significant improvement over traditional mapping methods. The study reveals an underestimation of organic soil coverage in Latvia by previous estimates, suggesting a broader distribution than recognized, with the model achieving an accuracy of 0.86 and a kappa value of 0.67. This research not only underscores the efficacy of integrating machine learning and remote sensing for soil mapping but also highlights the critical role of accurate data and models in determining organic soil distribution. The insights gained from this study are vital for policy-making and environmental planning, offering a more detailed understanding of Latviaʼs peatland resources and their conservation needs.