The derivation of global navigation satellite systems (GNSSs) tropospheric products is nowadays a state-of-the-art technique that serves both research and operational needs in a broad range of applications in meteorology. In particular, GNSS zenith tropospheric delay (ZTD) data assimilation is widely applied in Europe to enhance numerical weather predictions (NWPs). The current study presents the first attempt at introducing assimilation of ZTDs, derived from more than 48 stations of the Hellenic GNSS network, into the operational NWP system of the National Observatory of Athens (NOA) in Greece, which is based on the mesoscale Weather Research and Forecasting (WRF) model. WRF was applied during seven high-impact precipitation events covering the dry and wet season of 2018. The simulation employing the ZTD data assimilation reproduces more accurately, compared to the control experiment, the observed heavy rainfall (especially for high precipitation events, exceeding 20 mm in 24h) during both dry and wet periods. Assimilating ZTDs also improves the simulation of intense (>20 mm) convective precipitation during the time window of its occurrence in the dry season, and provides a beneficial influence during synoptic-scale events in the wet period. The above results, which are statistically significant, highlight an important positive impact of ZTD assimilation on the model’s precipitation forecast skill over Greece. Overall, the modelling system’s configuration, including the assimilation of ZTD observations, satisfactorily captures the spatial and temporal distribution of the observed rainfall and can therefore be used as the basis for examining further improvements in the future.

Image processing techniques that involve multispectral remotely sensed data are considered attractive for bathymetry applications as they provide a time- and cost-effective solution to water depths estimation. In this paper the potential of 8-bands image acquired by Worldview-2 satellite in providing precise depth measurements was investigated. Multispectral image information was integrated with available echo sounding and GPS data for the determination of the depth in the area of interest. In particular the main objective of this research was to evaluate the effectiveness of high spatial and spectral resolution of the new imagery data on water depth measurements using the Lyzenga linear bathymetry model. The existence of sea grass in a part of the study area influenced the linear relationship between water reflectance and depth. Therefore the bathymetric model was applied in three image parts: an area with sea grass, a mixed area and a sea grass-free area. In the last two areas the model worked successfully supported by the multiplicity of the imagery bands.

This paper presents a bee-condition-monitoring system incorporated with a deep-learning process to detect bee swarming. This system includes easy-to-use image acquisition and various end node approaches for either on-site or cloud-based mechanisms. This system also incorporates a new smart CNN engine called Swarm-engine for detecting bees and the issue of notifications in cases of bee swarming conditions to the apiarists. First, this paper presents the authors’ proposed implementation system architecture and end node versions that put it to the test. Then, several pre-trained networks of the authors’ proposed CNN Swarm-engine were also validated to detect bee-clustering events that may lead to swarming. Finally, their accuracy and performance towards detection were evaluated using both cloud cores and embedded ARM devices on parts of the system’s different end-node implementations.

The Strymon basin (Northern Greece) belongs to the geodynamically active regime of the Aegean and, as expected, it hosts active faults. Nevertheless, the study area exhibits a low instrumentally and historically recorded seismicity. In order to comprehend the crustal deformation, we implemented GNSS- and InSAR-based techniques. Global Navigation Satellite System (GNSS) primary geodetic data recorded by 32 permanent stations over 7 years were analyzed and input in the triangulation methodology so as to calculate a series of deformational parameters. Moreover, a geostatistical methodology indicated the spatial distribution of each parameter, showing strain delimited up to 2750 × 10⁻⁹. These results are in broad agreement with palaeoseismological surveys and active fault mapping. Moreover, InSAR analysis, based on a 6-year data recording, concluded that no horizontal rates have been traced in the E–W direction; if they do exist, they would be below resolution (less than 2 mm/yr). Peak vertical subsidence values of a few mm/yr are traced towards the hanging wall of the Serres fault zone within the Quaternary sediments at the eastern margin of Strymon basin but are attributed mainly to groundwater extraction. However, it is noteworthy that geodetic strain analysis implies: (a) that a couple of areas need further study to trace potentially active faults by palaeoseismological means; (b) the fault trace of the Serres fault zone might be further prolonged 8–10 km eastwards, where Quaternary sediments cover the fault.

In viticulture, downy mildew is one of the most common diseases that, if not adequately treated, can diminish production yield. However, the uncontrolled use of pesticides to alleviate its occurrence can pose significant risks for farmers, consumers, and the environment. This paper presents a new framework for the early detection and estimation of the mildew’s appearance in viticulture fields. The framework utilizes a protocol for the real-time acquisition of drones’ high-resolution RGB images and a cloud-docker-based video or image inference process using object detection CNN models. The authors implemented their framework proposition using open-source tools and experimented with their proposed implementation on the debina grape variety in Zitsa, Greece, during downy mildew outbursts. The authors present evaluation results of deep learning Faster R-CNN object detection models trained on their downy mildew annotated dataset, using the different object classifiers of VGG16, ViTDet, MobileNetV3, EfficientNet, SqueezeNet, and ResNet. The authors compare Faster R-CNN and YOLO object detectors in terms of accuracy and speed. From their experimentation, the embedded device model ViTDet showed the worst accuracy results compared to the fast inferences of YOLOv8, while MobileNetV3 significantly outperformed YOLOv8 in terms of both accuracy and speed. Regarding cloud inferences, large ResNet models performed well in terms of accuracy, while YOLOv5 faster inferences presented significant object classification losses.

In this study, we show how the Global Navigation Satellite System (GNSS)-derived vertical velocities contribute to the correction of tide gauge (TG) measurements used for the sea level rise estimation in Greece. Twelve sites with records of local sea level heights are processed in order to estimate their trend. Certain error sources related to TGs, e.g. equipment changes, data noise, may lead to biased or erroneous estimations of the sea level height. Therefore, it would be preferred to follow a robust estimation technique in order to detect and reduce outlier effects. The geocentric sea level rise is estimated by taking into account the land vertical motion of co-located GNSS permanent stations at the Hellenic area. TGs measure the height of the water relative to a monitored geodetic benchmark on land. On the other hand, using GNSS-based methods the vertical land motion can be derived. By means of extended models fitted to the GNSS time-series position, obtained from seven years of continuous data analysis, periodic signals are well described. The synergy of the two co-located techniques results in the correction of TG relative sea level heights taking into account the GNSS vertical velocities and consequently obtaining the conversion to absolute (geocentric) sea level trend.