Urban areas, characterized by dense populations and many socio-economic activities, increasingly suffer from floods, droughts, and heat stress due to land use and climate change. Traditionally, the urban thermal environment and water resources management have been studied separately, using urban land surface models (ULSMs) and urban hydrological models (UHMs). However, as our understanding deepens and the urgency to address future climate disasters grows, it becomes clear that hydrological disasters—such as floods, droughts, severe urban thermal environments, and more frequent heat waves—are actually not isolated events but compound events. This underscores the close interaction between the water cycle and the energy balance. Consequently, the existing separation between ULSMs and UHMs creates significant obstacles to better understanding urban hydrological and meteorological processes, which is crucial for addressing the high risks posed by climate change. Defining the future direction of process-based models for hydro-meteorological predictions and assessments is essential for better managing climate disasters and evaluating response measures in densely populated urban areas. Our review focuses on three critical aspects of urban hydro-meteorological simulation: similarities, differences, and gaps among different models; existing gaps in physical process implementations; and efforts, challenges, and potential for model coupling and integration. We find that ULSMs inadequately represent water surfaces and hydraulic systems, while UHMs lack explicit surface energy balance solutions and detailed building representations. Coupled models show potential for simulating urban hydro-meteorological environments, but face challenges at regional and neighborhood scales. Our review highlights the need for interdisciplinary communication between the urban climatology and urban water management communities to enhance urban hydro-meteorological simulation models.@en

In July 2021, extreme precipitation caused devastating flooding in Germany, Belgium and the Netherlands, particularly in the Geul River catchment. Such precipitation extremes had not been previously recorded and were not expected to occur in summer. This contributed to poor flood forecasting and, hence, extensive damage. Climate change was mentioned as a potential explanation for these unprecedented events. However, before such a statement can be made, we need a better understanding of the drivers of floods in the Geul and their long-Term variability, which are poorly understood and have not been recently examined. In this paper, we use an event-based approach to identify the dominant flood drivers in the Geul. We also employ (1) a multi-Temporal trend analysis to investigate their temporal variability and (2) a novel methodology to detect the dominant direction of any trend. Results suggest that extreme 24 h precipitation alone is typically insufficient to cause floods. The joint probability of extreme and prolonged rainfall combined with wet initial conditions (compound event) determines the chances of flooding. Flood-producing precipitation shows a consistent increase in the winter half-year, a period in which more than 70 % of extremely high flows have historically occurred. While no consistent trend patterns are evident in the majority of precipitation and extreme flow trends in the summer half-year, an increasing direction is visible in the recent past.

@en

Machine learning (ML) models offer advantages over process-based models for real-time reservoir operation modelling, yet the impact of input variable selection (IVS) and data pre-processing on model performance remains underexplored. This study investigates various input variables for simulating daily reservoir outflow, using the Sirikit reservoir in Thailand as a case study. The datasets include daily Sirikit storage and inflow, outflow of Bhumibol (neighbouring reservoir), downstream discharge, and temporal factors (month and day of the week). Time series decomposition and correlation analyses were used to assess data relationships. We tested seven ML models: multiple linear regression, support vector machine, K-nearest neighbour, classification and regression tree, random forest, multi-layer perceptron, and recurrent neural network (RNN). The optimal input set comprised the previous day’s storage, inflow from 2 days before to 2 days after, and month. With these inputs, all ML models simulated outflow adequately (KGE_training = 0.42–1.0 and KGE_testing = 0.46–0.56), with RNN showing the most potential for improvement. Input scaling significantly enhanced model performance, reducing RMSE_training by 44 m³ s^-1 and RMSE_testing by 14 m³ s^-1. This study’s novelty lies in its comprehensive insights of IVS and data scaling, highlighting their critical roles in enhancing ML model application for operational reservoir simulations.

@en

Potentially, the greatest benefit of commercial microwave links (CMLs) as opportunistic rainfall sensors lies in regions that lack dedicated rainfall sensors, most notably low-and middle-income countries. However, current CML rainfall retrieval algorithms are predominantly tuned and applied to (European) CML networks in temperate or Mediter-ranean climates. This study investigates whether local quantitative precipitation estimates from CMLs in a tropical region, specifically Sri Lanka, can be improved by optimizing two dominant parameters in the rainfall retrieval algorithm RAINLINK, namely, the wet antenna attenuation correction factor A_a and the relative contribution of minimum and max-imum received signal levels a. Using a grid search, based on 10 months of CML data from 22 link–gauge clusters consisting of 105 sublinks that lie within 1 km of a daily rain gauge, the optimal values of A_a and a are first derived for the entire country and compared to the default RAINLINK values. Subsequently, the CMLs are grouped by link length, frequency, climate zone, and daily rainfall depth classes, and A_a and a are derived for each of these classes. Calibrating parameters on all clusters across the country only leads to minor improvements. The actual optimal A_a and a values depend on the perfor-mance metric favored. Calibrating on network properties, particularly short link length and high-frequency classes, does significantly improve rainfall estimates. By relating the optimal A_a and a values to known network metadata, the results from this study are potentially applicable to other tropical CML networks that lack nearby reference rainfall data.

@en

Plastic is an emerging pollutant, and the quantities in rivers and oceans are expected to increase. Rivers are assumed to transport land-based plastic into the ocean, and the fluvial and marine transport processes have been relatively well studied to date. However, the processes controlling the transport in tidal rivers and estuaries, the interface between fluvial and marine systems, remain largely unresolved. For this reason, current estimates of riverine plastic pollution and export into the ocean remain highly uncertain. Hydrodynamics in tidal rivers and estuaries are influenced by tides and freshwater discharge. As a consequence, flow velocity direction and magnitude can change diurnally. In turn, this impacts the transport dynamics of solutes and pollutants, including plastics. Plastic transport dynamics in tidal rivers and estuaries remain understudied, yet the available observations suggest that plastics can be retained here for long time periods, especially during periods of low net discharge. Additional factors such as riparian vegetation and riverbank characteristics, in combination with bi-directional flows and varying water levels, can lead to an even higher likelihood of long-term retention. Here, we provide a first observation-based estimate of net plastic transport on a daily timescale in tidal rivers. For this purpose, we developed a simple Eulerian approach using sub-hourly observations of plastic transport and discharge during full tidal cycles. We applied our method to the highly polluted Saigon River, Vietnam, throughout six full tidal cycles in May 2022. We show that the net plastic transport is about 20%-33% of the total plastic transport. We found that plastic transport and river discharge are positively and significantly correlated (Pearson's R2Combining double low line0.76). The net transport of plastic is higher than the net discharge (20%-33% and 16%, respectively), suggesting that plastic transport is governed by factors other than water flow. Such factors include wind, varying plastic concentrations in the water, and entrapment of plastics downstream of the measurement site. The plastic net transport rates alternate between positive (seaward) net transport and negative (landward) net transport as a result of the diurnal inequality in the tidal cycles. We found that soft and neutrally buoyant items had considerably lower net transport rates than rigid and highly buoyant items (10%-16% vs. 30%-38%), suggesting that transport dynamics strongly depend on item characteristics. Our results demonstrate the crucial role of tidal dynamics and bi-directional flows in plastic transport dynamics. With this paper we emphasize the importance of understanding fundamental transport dynamics in tidal rivers and estuaries to ultimately reduce the uncertainties of plastic emission estimates into the ocean.

@en

The Goddard Profiling algorithm (GPROF) converts radiometer observations from Global Precipitation Measurement (GPM) constellation satellites into precipitation estimates. Typically, high-quality ground-based estimates serve as reference to evaluate GPROF's performance. To provide a fair comparison, the ground-based estimates are often spatially aligned to GPROF. However, GPROF combines observations from various sensors and channels, each associated with a distinct footprint. Consequently, uncertainties related to the representativeness of the sampled areas are introduced in addition to the uncertainty when converting brightness temperatures into precipitation intensities. The exact contribution of resampling precipitation estimates, required to spatially and temporally align different resolutions when combining or comparing precipitation observations, to the overall uncertainty remains unknown. Here, we analyze the current performance of GPROF over the Netherlands during a 4-year period (2017-2020) while investigating the uncertainty related to sampling. The latter is done by simulating the reference precipitation as satellite footprints that vary in size, geometry, and applied weighting technique. Only GPROF estimates based on observations from the conical-scanning radiometers of the GPM constellation are used. The reference estimates are gauge-adjusted radar precipitation estimates from two ground-based weather radars from the Royal Netherlands Meteorological Institute (KNMI). Echo top heights (ETHs) retrieved from the same radars are used to classify the precipitation as shallow, medium, or deep. Spatial averaging methods (Gaussian weighting vs. arithmetic mean) minimally affect the magnitude of the precipitation estimates. Footprint size has a higher impact but cannot explain all discrepancies between the ground- and satellite-based estimates. Additionally, the discrepancies between GPROF and the reference are largest for low ETHs, while the relative bias between the different footprint sizes and implemented weighting methods increase with increasing ETHs. Lastly, our results do not show a clear difference between coastal and land simulations. We conclude that the uncertainty introduced by merging different channels and sensors cannot fully explain the discrepancies between satellite- and ground-based precipitation estimates. Hence, uncertainties related to the retrieval algorithm and environmental conditions are found to be more prominent than resampling uncertainties, in particular for shallow and light precipitation.

@en

Radar rainfall nowcasting has mostly been applied to relatively large (often rural) domains (e.g., river basins), although rainfall nowcasting in small urban areas is expected to be more challenging. Here, we selected 80 events with high rainfall intensities (at least one 1-km ² grid cell experiences precipitation >15 mm h ²¹ for 1-h events or 30 mm day ²¹ for 24-h events) in five urban areas (Maastricht, Eindhoven, The Hague, Amsterdam, and Groningen) in the Netherlands. We evaluated the performance of 9060 probabilistic nowcasts with 20 ensemble members by applying the short-term ensemble prediction system (STEPS) from Pysteps to every 10-min issue time for the selected events. We found that nowcast errors increased with decreasing (urban) areas especially when below 100 km ². In addition, at 30-min lead time, the underestimation of nowcasts was 38% larger and the discrimination ability was 11% lower for 1-h events than for 24-h events. A set of gridded correction factors for the Netherlands, CARROTS (Climatology-based Adjustments for Radar Rainfall in an Operational Setting) could adjust the bias in real-time QPE and nowcasts by 70%. Yet, nowcasts were still found to underestimate rainfall more than 50% above 40-min lead time relative to the reference, which indicates that this error originates from the nowcasting model itself. Also, CARROTS did not adjust the rainfall spatial distribution in urban areas much. In summary, radar-based nowcasting for urban areas (between 67 and 213 km ²) in the Netherlands exhibits a short skillful lead time of about 20 min, which can only be used for last-minute warning and preparation.

@en

Spaceborne microwave radiometers represent an important component of the Global Precipitation Measurement (GPM) mission due to their frequent sampling of rain systems. Microwave radiometers measure microwave radiation (brightness temperatures Tb), which can be converted into precipitation estimates with appropriate assumptions. However, detecting shallow precipitation systems using spaceborne radiometers is challenging, especially over land, as their weak signals are hard to differentiate from those associated with dry conditions. This study uses a random forest (RF) model to classify microwave radiometer observations as dry, shallow, or nonshallow over the Netherlands}a regionwith varying surface conditions and frequent occurrence of shallow precipitation. The RF model is trained on five years of data (2016–20) and tested with two independent years (2015 and 2021). The observations are classified using ground-based weather radar echo top heights. Various RF models are assessed, such as using only GPM Microwave Imager (GMI) Tb values as input features or including spatially aligned ERA5 2-m temperature and freezing level reanalysis and/or Dual-Frequency Precipitation Radar (DPR) observations. Independent of the input features, the model performs best in summer and worst in winter. The model classifies observations from high-frequency channels ($85 GHz) with lower Tb values as nonshallow, higher values as dry, and those in between as shallow. Misclassified footprints exhibit radiometric characteristics corresponding to their assigned class. Case studies reveal dry observations misclassified as shallow are associated with lower Tb values, likely resulting from the presence of ice particles in nonprecipitating clouds. Shallow footprints misclassified as dry are likely related to the absence of ice particles. SIGNIFICANCE STATEMENT: Published research concerning rainfall retrieval algorithms from microwave radiometers is often focused on the accuracy of these algorithms. While shallow precipitation over land is often characterized as problematic in these studies, little progress has been made with these systems. In particular, precipitation formed by shallow clouds, where shallow refers to the clouds being close to Earth’s surface, is often missed. This study is focused on detecting shallow precipitation and its physical characteristics to further improve its detection from spaceborne sensors. As such, it contributes to understanding which shallow precipitation scenes are challenging to detect from microwave radiometers, suggesting possible ways for algorithm improvement.

@en

This chapter reviews the state-of-the-art of land surface rainfall estimation using measurements from weather radars, personal weather stations, and commercial microwave links, including comparisons to rain gauge measurements. These studies are related to recently emerging field of opportunistic weather sensing using the existing communications infrastructure.@en

Nowcasting leverages real-time atmospheric conditions to forecast weather over short periods. State-of-the-art models, including PySTEPS, encounter difficulties in accurately forecasting extreme weather events because of their unpredictable distribution patterns. In this study, we design a physics-informed neural network to perform precipitation nowcasting using the precipitation and meteorological data from the Royal Netherlands Meteorological Institute (KNMI). This model draws inspiration from the novel Physics-Informed Discriminator GAN (PID-GAN) formulation, directly integrating physics-based supervision within the adversarial learning framework. The proposed model adopts a GAN structure, featuring a Vector Quantization Generative Adversarial Network (VQ-GAN) and a Transformer as the generator, with a temporal discriminator serving as the discriminator. Our findings demonstrate that the PID-GAN model outperforms numerical and SOTA deep generative models in terms of precipitation nowcasting downstream metrics.

@en

Hydrometeorological processes are often assumed to be key drivers of plastic transport. However, the predominant focus on these factors overlooks the impact of anthropogenic factors, such as mismanaged plastic waste (MPW) on plastic transport variability. Here, we investigate the roles of both anthropogenic and hydrometeorological factors on plastic pollution in the Odaw catchment, Ghana. Data on macroplastic transport and density were collected at ten locations between December 2021 and December 2022. We tested for differences between the wet and dry seasons and applied a multiple regression analysis to examine the separate and combined impact of hydrometeorological variables (rainfall, discharge, and windspeed) on macroplastic transport. Additionally, we analyzed the spatial correlation in macroplastic transport/density with MPW and population density. Data collection involved visual counting of floating macroplastics at 10 river locations and counting litter at 9 riverbanks and land locations. Rainfall data was sourced from TAHMO (Trans-African Hydrometeorological Observatory), discharge was measured during field campaigns, and windspeed data sourced from a global climate data provider. We used globally modelled MPW estimates to represent anthropogenic factors. Contrary to previous studies, we found no seasonal differences in macroplastic pollution and only weak correlations were observed between the hydrometeorological variables and macroplastic transport. However, a strong correlation was observed between MPW and macroplastic pollution. We hypothesize that, the influence of hydrometeorological factors on macroplastic transport depend on the relative impact of anthropogenic factors. Our research highlights the limited role of hydrometeorology, showing the significant role of mismanaged plastic waste to field monitored macroplastic pollution variability in the catchment. This insight is essential for future research as it highlights the importance of holistically investigating both anthropogenic and hydrometeorological factors in explaining plastic transport and retention dynamics. This insight is essential for developing interventions that effectively address plastic pollution in catchments.

@en

Terrestrial microwave links are increasingly being used to estimate path-averaged precipitation by determining the attenuation caused by rainfall along the link path, mostly with commercial microwave links from cellular telecommunication networks. However, the temporal resolution of these rainfall estimates and the method to derive them are often determined by the temporal sampling strategy that is employed by the mobile network operators. Currently, the links are most often sampled at a temporal resolution of 15 min with a recording of the minimum and maximum values, while more recently, a form of instantaneous sampling with possible intervals up to 1 s has also been set up. For rainfall research purposes, often high temporal resolutions in combination with averaged values are preferred. However, it is uncertain how these various temporal sampling strategies affect the estimated rainfall intensity. Here we aim to understand how temporal sampling strategies affect the measured rainfall intensities using microwave links. To do so, we use data from three collocated microwave links, two 38 GHz and one 26 GHz, sampled at 20 Hz and covering a 2.2 km path over the city of Wageningen, the Netherlands. We aggregate the microwave link power levels to multiple time intervals (1 s to 60 min) and use a mean, instantaneous, and minimum and maximum value to characterize the signal. Based on the aggregated data, we compute rainfall intensities and compare these with 20 Hz rainfall estimates, such that we isolate errors and uncertainties caused by the sampling strategies from instrumental effects, such as different biases between instruments and representativeness errors. In general, our results show that for all sampling strategies, an increase in sampling time interval reduces the performance of the rainfall estimates, which especially holds for the instantaneous sampling strategy. Even the mean sampling strategy, which generally performs best of all strategies, is sensitive to this reduction in temporal resolution and could lead to significant underestimations. This sensitivity of the mean sampling to the temporal resolution seems to be largely affected by the non-linear relation between attenuation and rainfall. The min–max sampling strategy is mostly prone to minor underestimations or large overestimations of the path-averaged rainfall intensities. Moreover, our results, including a comparison with theoretical events, show that the attenuation due to wet antennas not only affects the comparison between the rainfall estimates obtained with a microwave link and another reference instrument but also has a significant influence on the performance of the rainfall retrieval algorithm, especially for devices with relatively long duration of the wet-antenna attenuation combined with the longer sampling time intervals. Overall, this study demonstrates the effect a selected sampling strategy can have on rainfall intensity estimates using (commercial) microwave links.@en

Rivers are one of the main conduits that deliver plastic from land into the sea, and also act as reservoirs for plastic retention. Yet, our understanding of the extent of river exposure to plastic pollution remains limited. In particular, there has been no comprehensive quantification of the contributions from different river compartments, such as the water surface, water column, riverbank and floodplain to the overall river plastic transport and storage. This study aims to provide an initial quantification of these contributions. We first identified the main relevant transport processes for each river compartment considered. We then estimated the transport and storage terms, by harmonizing available observations on surface, suspended and floodplain plastic. We applied our approach to two river sections in The Netherlands, with a focus on macroplastics (≥2.5 cm). Our analysis revealed that for the studied river sections, suspended plastics account for over 96% of item transport within the river channel, while their relative contribution to mass transport is only 30%–37% (depending on the river section considered). Surface plastics predominantly consisted of heavier items (mean mass: 7.1 g/#), whereas suspended plastics were dominated by lighter fragments (mean mass: 0.1 g/#). Additionally, the majority (98%) of plastic mass was stored within the floodplains, with the river channel accounting for only 2% of the total storage. Our study developed a harmonized approach for quantifying plastic transport and storage across different river compartments, providing a replicable methodology applicable to different regions. Our findings emphasize the importance of systematic monitoring programs across river compartments for comprehensive insights into riverine plastic pollution.

@en

The number of rainfall observations from personal weather stations (PWSs) has increased significantly over the past years; however, there are persistent questions about data quality. In this paper, we reflect on three quality control algorithms (PWSQC, PWS-pyQC, and GSDR-QC) designed for the quality control (QC) of rainfall data. Technical and operational guidelines are provided to help interested users in finding the most appropriate QC to apply for their use case. All three algorithms can be accessed within the OpenSense sandbox where users can run the code. The results show that all three algorithms improve PWS data quality when cross-referenced against a rain radar data product. The considered algorithms have different strengths and weaknesses depending on the PWS and official data availability, making it inadvisable to recommend one over another without carefully considering the specific setting. The authors highlight a need for further objective quantitative benchmarking of QC algorithms. This requires freely available test datasets representing a range of environments, gauge densities, and weather patterns.

@en

In July 2021, extreme precipitation caused devastating flooding in Germany, Belgium and the Netherlands, particularly in the Geul River catchment. Such precipitation extremes had not been previously recorded and were not expected to occur in summer. This contributed to poor flood forecasting and, hence, extensive damage. Climate change was mentioned as a potential explanation for these unprecedented events. However, before such a statement can be made, we need a better understanding of the drivers of floods in the Geul and their long-term variability, which are poorly understood and have not been recently examined. In this paper, we use an event-based approach to identify the dominant flood drivers in the Geul. We also employ (1) a multi-temporal trend analysis to investigate their temporal variability and (2) a novel methodology to detect the dominant direction of any trend. Results suggest that extreme 24 h precipitation alone is typically insufficient to cause floods. The joint probability of extreme and prolonged rainfall combined with wet initial conditions (compound event) determines the chances of flooding. Flood-producing precipitation shows a consistent increase in the winter half-year, a period in which more than 70 % of extremely high flows have historically occurred. While no consistent trend patterns are evident in the majority of precipitation and extreme flow trends in the summer half-year, an increasing direction is visible in the recent past.@en

Rivers represent one of the main conduits for the delivery of plastics to the sea, while also functioning as reservoirs for plastic retention. In tropical regions, rivers are exposed to both high levels of plastic pollution and invasion of water hyacinths. This aquatic plant forms dense patches at the river surface that drift due to winds and currents. Recent work suggests that water hyacinths play a crucial role in influencing plastic transport, by efficiently trapping the majority of surface plastic within their patches. However, a comprehensive understanding of the interaction between water hyacinths and plastics is still lacking. We hypothesize that the properties relevant to plastic transport change due to their trapping in water hyacinth patches. In particular, the length scale, defined as the characteristic size of the transported material, is a key property in understanding how materials move within rivers. Here, we show that water hyacinth patches trap on average 54%–77% of all observed surface plastics at the measurement site (Saigon river, Vietnam). Both temporally and spatially, we found that plastic and water hyacinth presence co-occur. The formation of plastic-plant aggregates carries significant implications for both clean-up and monitoring purposes, as these aggregates can be detected from space and need to be jointly removed. In addition, the length scale of trapped plastics (∼4.0 m) was found to be forty times larger than that of open water plastics (∼0.1 m). The implications of this increased length scale for plastic transport dynamics are yet to be fully understood, calling for further investigation into travel distances and trajectories. The effects of plastic trapping likely extend to other key properties of plastic-plant aggregates, such as effective buoyancy and mass. Given the prevalence of plant invasion and plastic pollution in rivers worldwide, this research offers valuable insights into the complex environmental challenges faced by numerous rivers.

@en

Machine learning (ML) models offer advantages over process-based models for real-time reservoir operation modelling, yet the impact of input variable selection (IVS) and data pre-processing on model performance remains underexplored. This study investigates various input variables for simulating daily reservoir outflow, using the Sirikit reservoir in Thailand as a case study. The datasets include daily Sirikit storage and inflow, outflow of Bhumibol (neighbouring reservoir), downstream discharge, and temporal factors (month and day of the week). Time series decomposition and correlation analyses were used to assess data relationships. We tested seven ML models: multiple linear regression, support vector machine, K-nearest neighbour, classification and regression tree, random forest, multi-layer perceptron, and recurrent neural network (RNN). The optimal input set comprised the previous day’s storage, inflow from 2 days before to 2 days after, and month. With these inputs, all ML models simulated outflow adequately (KGEtraining = 0.42–1.0 and KGEtesting = 0.46–0.56), with RNN showing the most potential for improvement. Input scaling significantly enhanced model performance, reducing RMSEtraining by 44 m3 s-1 and RMSEtesting by 14 m3 s-1. This study’s novelty lies in its comprehensive insights of IVS and data scaling, highlighting their critical roles in enhancing ML model application for operational reservoir simulations.@en

Plastic pollution in rivers threatens ecosystems, increases flood risk due to its accumulations at hydraulic structures and its final emissions into the ocean threaten aquatic life, especially and probably most in coastal urbanized areas. Previous work suggests that plastic pollution in these urban rivers is influenced by hydrometeorological and anthropogenic factors. However, the transport dynamics of the plastics in such rivers are non-linear and complex and remain largely unresolved. Here, we show that tidal dynamics can be the main driver of plastic transport closest to the river mouth. Outside the tidal zone, rainfall and river discharge were identified to be more important drivers. We monitored plastic transport in the Odaw river, Ghana during the dry season. The Odaw drains the densely populated city of Accra and discharges into the Gulf of Guinea. Data were collected between March and May 2021 (dry season), using visual counting at four bridges along the river, of which two were located within the tidal zone. We explored the correlations between river plastic transport, and rainfall, tidal dynamics, and river discharge. Finally, we estimated the total plastic mass transport by using item-to-mass conversion data from previously published literature. We observed a peak in plastic transport at the upstream bridge within the tidal zone after an increase in rainfall (7.3 times larger). We found a gradient of the hydrometeorological factors driving plastic transport. Closer to the river mouth, tidal dynamics were more strongly correlated with plastic transport than upstream. The daily mass transport was estimated to be between 1.4–3.8 × 10² kg/d, which is lower than previous model estimates. These results add to the evidence of inconsistent correlations between plastic transport and hydrometeorological variables. Long-term monitoring data is required to further investigate this. The results also support the hypothesis that tidal dynamics are a crucial factor in controlling the emissions of plastics from rivers into the ocean. The findings provide a baseline for the Odaw river during the dry season and allow for comparison with the wet season. The approach adopted here also serves as a blueprint for similar urban river systems, regionally and globally.

@en

Accurate precipitation observations are crucial for hydrological forecasts, notably over rapidly responding urban areas. This study evaluated the accuracy of three gridded spaceborne rainfall products (Integrated Multi-satellitE Retrievals for GPM (IMERG), Meteosat Second Generation Visible (MSG-VIS), and MSG-Infrared (MSG-IR)) and the non-governmental Trans-African Hydro-Meteorological Observatory (TAHMO) gauges across the Odaw catchment (Accra, Ghana) from January 2020-July 2022. IMERG is hardly able to capture the strong spatial variability of rainfall required for flood forecasting, but agrees in annual sums with TAHMO and MSG-IR. MSG-IR has difficulties during the wet season. MSG-VIS, only available during daylight, shows limited accuracy and gives high estimates while other products do not detect rain. TAHMO gauges effectively record high-intensity events and their strong spatial variability, although some (daily) accumulations are doubtful and data gaps exist due to technical issues. These findings assist hydrological modelers in selecting appropriate datasets at suitable spatiotemporal resolutions for their research.

@en

Nowcasting is an observation-based method that uses the current state of the atmosphere to forecast future weather conditions over several hours. Recent studies have shown the promising potential of using deep learning models for precipitation nowcasting. In this paper, novel deep generative models are proposed for precipitation nowcasting. These models are equipped with extreme-value losses to more reliably predict extreme precipitation events. The proposed deep generative model contains a Vector Quantization Generative Adversarial Network and a Transformer ("VQGAN + Transformer"). For enhanced modeling and forecasting of extreme events, Extreme Value Loss (EVL) is incorporated in the autore-gressive Transformer. The numerical results show that the proposed model achieves comparable performance with the state-of-the-art conventional nowcasting method PySTEPS for predicting nominal values. By incorporating an EVL, the proposed model yields more accurate nowcasting of extreme precipitation.@en