High-resolution Spatiotemporal Data Research Articles

Full-coverage estimation of CO2 concentrations in China via multisource satellite data and Deep Forest model.

Monitoring China's carbon dioxide (CO2) concentration is essential for formulating effective carbon cycle policies to achieve carbon peaking and neutrality. Despite insufficient satellite observation coverage, this study utilizes high-resolution spatiotemporal data from the Orbiting Carbon Observatory 2 (OCO-2), supplemented with various auxiliary datasets, to estimate full-coverage, monthly, column-averaged carbon dioxide (XCO2) values across China from 2015 to 2022 at a spatial resolution of 0.05° via the deep forest model. The 10-fold cross-validation results indicate a correlation coefficient (R) of 0.95 and a determination coefficient (R²) of 0.90. Validation against ground-based station data yielded R values of 0.93, and R² values reached 0.81. Further validation from the Greenhouse Gases Observing Satellite (GOSAT) and the Copernicus Atmosphere Monitoring Service Reanalysis dataset (CAMS) produced R² values of 0.87 and 0.80, respectively. During the study period, CO2 concentrations in China were higher in spring and winter than in summer and autumn, indicating a clear annual increase. The estimates generated by this study could potentially support CO2 monitoring in China.

A Radar-Based Methodology for Aerosol Plume Identification and Characterisation on the South African Highveld

Biomass burning on the South African Highveld annually injects substantial amounts of aerosols and trace gases into the atmosphere, impacting the global radiative balance, cloud microphysics, and regional air quality. These aerosols are transported as plumes over long distances, posing challenges to existing in situ and satellite-based monitoring techniques because of their limited spatial and temporal resolution, particularly in environments with low-level sources. This study aims to develop and validate a novel radar-based methodology to detect, track, and characterise aerosol plumes, addressing the limitations of existing in situ and satellite monitoring techniques. Using high-resolution volumetric reflectivity data from an S-band radar in Pretoria, South Africa, a traditional storm tracking algorithm is adapted to improve plume identification. Case studies of plume events in June and August 2013 demonstrate the radar’s effectiveness in distinguishing lower vertical profiles and reduced reflectivity of plumes compared with storm echoes. The adapted algorithm successfully tracked the spatial and temporal evolution of the plumes, revealing their short-lived nature. Results indicate that radar-derived geospatial characteristics have the potential to contribute significantly to understanding the impacts of plumes on local air quality. These findings underscore the critical need for high spatio-temporal resolution data to support effective air quality management and inform policy development in regions affected by biomass burning.

Exploring the effects of climate change and urban policies on lake water quality using remote sensing and explainable artificial intelligence

Climate change and urban policies significantly impact lake water quality, especially in rapidly developing regions. Existing impact assessments are limited by the duration and resolution of water quality data, while long-term impacts and contributions of climate and socioeconomic drivers have rarely been quantified. In this study, we utilized remote sensing imagery to obtain high spatiotemporal resolution data and Explainable Artificial Intelligence (XAI) to reveal how the comprehensive impacts vary over time. We obtained Trophic State Index (TSI) data between 2000 and 2022 from remote sensing imagery. XAI was employed to elucidate the intricate connections between TSI and the multifaceted factors that encompass climate and socioeconomic dynamics. Six typical lakes in East China were selected for this study. Results showed that the TSI of these lakes had generally decreased over the past two decades, being typically higher in the littoral zones compared to the center. Moreover, higher temperatures increased lake TSI. Meanwhile, higher rainfall increased TSI through increased lake nutrient inputs from surrounding watersheds where agricultural and industrial activities are intensive. It was also found that economic development impacts on lake TSI shifted around 2013. Economic development caused severe negative effects on lake water quality before 2013, while this phenomenon decreased thereafter. This is attributed to the implementation of a series of green transformation policies in the region. This study demonstrates the long-term contribution of climate change and urban policies in driving lake water quality variations, and provides implications for policy makers to establish positive economic, social, and environmental linkages.

Exploring spatiotemporal dynamics, seasonality, and time-of-day trends of PM2.5 pollution with a low-cost sensor network: Insights from classic and spatially explicit Markov chains

Fine particulate matter (PM2.5) is a major health and environmental concern, with significant spatiotemporal dynamics in urban areas. Low-cost air quality sensor (LCS) networks offer a paradigm-changing opportunity to acquire high spatiotemporal resolution data, revealing the urban pollution landscape with sufficient detail for effective policymaking and health assessment. This study advances geospatial air quality research by using classic and spatial Markov chains to analyze the seasonality and intra-daily variations of PM2.5 using LCS data. Results highlight distinctive PM2.5 seasonality, with the “Good” state predominating in summer and being least common in winter. Midday is the peak period for the “Good” state, while mornings and nights have poorer conditions, suggesting a need for stricter pollution control during evening traffic rush hours. Notably, the impact of temporal scale on spatial Markov analysis is substantial, showing a broader range of air pollution states, increased stability, and reduced variation between time intervals compared to daily assessments. Site-level analysis reveals that rural sites are more likely to maintain “Good” state and less likely to transition out of it. Overall, this study highlights the effectiveness of high spatiotemporal resolution data and demonstrates the capacity of Markov chains to reveal nuances in phenomena such as air pollution.

Boiling crisis is a consequence of thermal runaway in the substrate due to degradation of nucleate boiling heat transfer

Chaotic bubble interactions during vigorous boiling have foiled attempts at an accurate mechanistic understanding of this important industrial transport process. Prediction of the boiling crisis (dryout) that occurs due to the spontaneous merger of bubbles into an insulating vapor film at critical heat flux remains an unsolved challenge. This work diagnoses the dryout process using synchronized high-resolution spatiotemporal heat flux and phase data, obtained via through-substrate infrared and visual inspection. We examine prevailing theories for the boiling crisis and provide evidence to rule out all but one. The boiling crisis is found to be a consequence of a peak in the nucleate boiling curve, past which the degradation of boiling heat transfer and concomitant increase in superheat are caused by replacement of the thermally efficient contact line region with the inefficient vapor-covered region for the surface-fluid combination studied. This results in substrate thermal runaway and dryout. Rather than seeking a separate dryout trigger mechanism, nucleate boiling models must instead inherently capture this peak. The heat flux partitioning employed demonstrates that critical heat flux can indeed be predicted by capturing this peak. A generalized framework is suggested for predicting the boiling curve as part of a multidimensional surface, a boiling manifold.

Assessing the spatial transferability of calibration models across a low-cost sensors network

Low-cost sensor networks (LCSNs) are expanding worldwide to gather high spatiotemporal resolution data due to their economic feasibility and compact size. The reliability of LCS-recorded data is limited due to their calibration dependencies in the field. Previous studies have focused on the development of LCS calibration models by co-location with the regulatory monitoring stations. However, it is challenging to calibrate LCS in the field for countries with limited infrastructure for air quality monitoring, pointing towards the need for transferable calibration models. Only a few studies have addressed this challenge and provide no information on the factors that may affect the performance of transferable calibration models. Here, we examined the spatial transferability of the calibration models developed using machine learning (ML) algorithms for an LCSN with twenty-two (22) sites in NCT-Delhi. The site-specific calibration models performed well at each site with high R2 and significantly low RMSE values. These models were transferred to the other sites, and the effect of distance between the sites (D), source composition, PM ratios, and particle size distribution (PSD) on the transferability of calibration models was investigated. The models developed at the Mundka (S10) and Punjabi Bagh (S16) sites complied with the evaluation criterion (R2 ≥ 0.70) for each site, irrespective of the distance between the sites. Furthermore, PM ratios reported by the LCSs did not significantly differ across sites, suggesting that the PMS algorithm provides a proxy of the size-resolved mass fractions. Evaluation of the PSD at different sites supported our findings. We also introduced the concept of selecting representative locations for LCS co-location by computing transferability scores using k-means clustering and presented a reference map for NCT-Delhi for developing scalable calibration models.

A sentinel watching over inter-tidal seagrass phenology across Western Europe and North Africa

Seagrasses are marine flowering plants that form extensive meadows from the inter-tidal zone up to ~50 m depth. As biological and ecological Essential Biodiversity Variables, seagrass cover and composition provide a wide range of ecosystem services. Inter-tidal seagrass meadows provide services to many ecosystems, so monitoring their occurrence, extent, condition and diversity can be used to indicate the biodiversity and health of local ecosystems. Current global estimates of seagrass extent and recent reviews either do not mention inter-tidal seagrasses and their seasonal variation, or combine them with sub-tidal seagrasses. Here, using high-spatial and high-temporal resolution satellite data (Sentinel-2), we demonstrate a method for consistently mapping inter-tidal seagrass meadows and their phenology at a continental scale. We were able to highlight varying seasonal patterns that are observable across a 23° latitudinal range. Timings of peaks in seagrass extent varied by up to 5 months, rather than the previously assumed marginal to non-existent variation in peak timing. These results will aid management by providing high-resolution spatio-temporal monitoring data to better inform seagrass conservation and restoration. They also highlight the high level of seasonal variability in inter-tidal seagrass, meaning combination with sub-tidal seagrass for global assessments will likely produce misleading or incorrect estimates.

Estimation of Near-surface Ozone Concentration in the Beijing-Tianjin-Hebei Region Based on XGBoost-LME Model

High spatiotemporal resolution data on near-surface ozone concentration distribution is of great significance for monitoring and controlling atmospheric ozone pollution and improving the living environment. Using TROPOMI-L3 NO2, HCHO products, and ERA5-land high-resolution data as estimation variables, an XGBoost-LME model was constructed to estimate the near-surface ozone concentration in the Beijing-Tianjin-Hebei Region. The results showed that： ① Through correlation analysis, surface 2 m temperature （T2M）, 2 m dewpoint temperature （D2M）, surface solar radiation downwards （SSRD）, tropospheric formaldehyde （HCHO）, and tropospheric nitrogen dioxide （NO2） were important factors affecting the near-surface ozone concentration in the Beijing-Tianjin-Hebei Region. Among them, T2M, SSRD, and D2M had strong correlations, with correlation coefficients of 0.82, 0.75, and 0.71, respectively. ② Compared with that of other models, the XGBoost-LME model had the best performance in terms of various indicators. The ten-fold cross-validation evaluation indicators R2, MAE, and RMSE were 0.951, 9.27 μg·m-3, and 13.49 μg·m-3, respectively. At the same time, the model performed well at different time scales. ③ In terms of time, there was a significant seasonal difference in near-surface ozone concentration in the Beijing-Tianjin-Hebei Region in 2019, with the concentration changing in the order of summer > spring > autumn > winter. The monthly average ozone concentration in the region showed an inverted "V" trend, with a slight increase in September. The highest value occurred in July, whereas the lowest value occurred in December. In terms of spatial distribution, the near-surface ozone concentrations in the Beijing-Tianjin-Hebei Region during the months of February and March were generally at the same levels. In January, November, and December, there was a relatively insignificant trend of higher concentrations in the north and lower concentrations in the south. For the remaining months, the spatial distribution of near-surface ozone concentrations in this area predominantly exhibited a pattern of higher concentrations in the south and lower concentrations in the north. High-value areas were predominantly found in the plain regions of the southern part with lower altitudes, dense population, and higher industrial emissions； low-value areas, on the other hand, were primarily located in mountainous areas of the northern part with higher altitudes, sparse population, higher vegetation coverage, and lower industrial emissions.

Feasibility of low-cost particulate matter sensors for long-term environmental monitoring: Field evaluation and calibration

Low-cost sensor networks offer the potential to reduce monitoring costs while providing high-resolution spatiotemporal data on pollutant levels. However, these sensors come with limitations, and many aspects of their field performance remain underexplored. During October to December 2023, this study deployed two identical low-cost sensor systems near an urban standard monitoring station to record PM2.5 and PM10 concentrations, along with environmental temperature and humidity. Our evaluation of the monitoring performance of these sensors revealed a broad data distribution with a systematic overestimation; this overestimation was more pronounced in PM10 readings. The sensors showed good consistency (R2 > 0.9, NRMSE<5 %), and normalization residuals were tracked to assess stability, which, despite occasional environmental influences, remained generally stable. A lateral comparison of four calibration models (MLR, SVR, RF, XGBoost) demonstrated superior performance of RF and XGBoost over others, particularly with RF showing enhanced effectiveness on the test set. SHAP analysis identified sensor readings as the most critical variable, underscoring their pivotal role in predictive modeling. Relative humidity consistently proved more significant than dew point and temperature, with higher RH levels typically having a positive impact on model outputs. The study indicates that, with appropriate calibration, sensors can supplement the sparse networks of regulatory-grade instruments, enabling dense neighborhood-scale monitoring and a better understanding of temporal air quality trends.

Advancing air quality monitoring: A low-cost sensor network in motion – Part I

In urban areas, high levels of air pollution pose significant risks to human health, emphasising the need for detailed air quality (AQ) monitoring. However, traditional AQ monitoring relies on the data from Reference Monitoring Stations, which are sparsely distributed and provide only hourly or daily data, failing to capture the spatial and temporal variability of air pollutant concentrations. Addressing this challenge, we introduce in this article the ExpoLIS system, an all-weather mobile AQ monitoring system that integrates various AQ low-cost sensors (LCSs), providing high spatio-temporal resolution data. This study demonstrates that the inclusion of an extended sampling device may mitigate the effect of the meteorological parameters and other disturbances on readings. At the same time, it did not reduce the quality of the data, both in static conditions and in motion, as we were able to maintain a certain level of agreement between the LCSs. In conclusion, the ExpoLIS system proves its versatility by enabling the collection of large quantities of accurate data, allowing a deeper understanding of the AQ dynamics in urban environments.

Comparison of IDW, Kriging and orographic based linear interpolations of rainfall in six rainfall regimes of Ethiopia

Study regionSix rainfall regimes of Ethiopia covering an area of 1.13 million km2. Study focusHydrological studies in mountainous terrain require high-resolution spatiotemporal rainfall data. Individual linear interpolation based on elevation, aspect and slope, multivariate linear interpolation,inverse distance weighted, and kriging were employed in six regimes. The performance of the six methods were assessed through cross validation of rain gauge stations and ENACTS data. New hydrological insightsWe adopted different coefficients for six rainfall regimes with respect to orographic liner interpolations. The results confirmed the importance of six rainfall regimes to carry out analysis of the interpolations. This is verified by the fact that performance of the products derived from the approaches differed from regime to regime. Elevation based interpolation is the dominant in regime I and IV. For regime II and III respectively slope and aspect based interpolations are performed well, whereas kriging interpolation is exhibited better in both regime V and VI. It is difficult to get a single best rainfall interpolation method for the Ethiopia heterogeneous land surface. Thus, this study recommends elevation based interpolation to be employed in the mountainous areas where as Kriging can be applied in the area where orographic and local climatic variability are insignificant. For interpolation of rainfall stations, it is essential to apply the methods after classify the study regions based on rainfall variability.

Heat exposure assessment based on high-resolution spatio-temporal data of population dynamics and temperature variations

Urban heat waves pose a significant risk to the health and safety of city dwellers, with urbanization potentially amplifying the health impact of extreme heat. Accurate assessments of population heat exposure hinge on the interplay between temperature, population spatial dynamics, and the epidemiological effects of temperature on health. Yet, many past studies have over-simplified the matter by assuming static populations, leading to substantial inaccuracies in heat exposure assessments. To address these issues, this study integrates dynamic population data, fluctuating temperature, and the exposure-response relationship between temperature and health to construct an advanced heat exposure assessment framework predicated on a population dynamic model. We analyzed urban heat island characteristics, population dynamics, and heat exposure during heat wave conditions in Beijing, a major city in China. Our findings highlight significant intra-day population movement between urban and suburban areas during heat wave conditions, with spatial population flow patterns showing clear scale-dependent characteristics. These population flow dynamics intensify heat exposure levels, and the disparity between dynamic population-weighted temperature and average temperature is most pronounced at night. Our research provides a more comprehensive understanding of real urban population heat exposure levels and can furnish city administrators with more scientifically rigorous evidence.

Inequalities in experiencing urban functions. An exploration of human digital (geo-)footprints

Studies on mobility inequalities have so far mostly relied on Survey data or Censuses. While such studies have demonstrated that inequalities strongly influence everyday mobility choices, these data sources lack granular information on people’s movements on a daily basis. By capitalising on high spatio-temporal resolution data provided by Spectus.ai, this study aims at investigating how the deprivation level of the area where people live influences the kinds of urban environment they are more likely to use for their everyday activities. To do this, raw GPS trajectories collected in 2019 in Great Britain (GB) are transformed into semantic trajectories where short-time changes and the functional nature of urban contexts are acknowledged as two key dimensions to understand human spatial behaviours. Hourly sequences of stops are extracted from GPS trajectories and enriched with contextual information based on a new area-based classification detecting urban functions. The data exploration shows that some human patterns are widely common across all levels of deprivation, such as the tendency to be mostly exposed to the urban context near the home location. At the same time, we show that differences exist, especially between those who live in the most deprived areas and those who live in the least deprived areas of GB. It appears that people living in the most deprived areas tend to have a less regular working pattern and be more exposed to urban-based functions and well-served areas, while those living in the least deprived areas have a more regular working patterns and are mostly exposed to the countryside and low-density areas. Our approach and results provide new insights on the temporal and contextual dimensions of mobility inequalities, informing on who is exposed to issues characterising certain urban environments.

Integrating LEO and GEO Observations: Toward Optimal Summertime Satellite Precipitation Retrieval

Abstract Reliable quantitative precipitation estimation with a rich spatiotemporal resolution is vital for understanding the Earth’s hydrological cycle. Precipitation estimation over land and coastal regions is necessary for addressing the high degree of spatial heterogeneity of water availability and demand, and for resolving the extremes that modulate and amplify hazards such as flooding and landslides. Advancements in computation power along with unique high spatiotemporal and spectral resolution data streams from passive meteorological sensors aboard geosynchronous Earth-orbiting (GEO) and low Earth-orbiting (LEO) satellites offer exciting opportunities to retrieve information about surface precipitation phenomena using data-driven machine learning techniques. In this study, the capabilities of U-Net–like architecture are investigated to map instantaneous, summertime surface precipitation intensity at the spatial resolution of 2 km. The calibrated brightness temperature products from the Global Precipitation Measurement (GPM) Microwave Imager (GMI) radiometer are combined with multispectral images (visible, near-infrared, and infrared bands) from the Advanced Baseline Imager (ABI) aboard the GOES-R satellites as main inputs to the U-Net–like precipitation algorithm. Total precipitable water and 2-m temperature from the Global Forecast System (GFS) model are also used as auxiliary inputs to the model. The results show that the U-Net–like algorithm can capture fine-scale patterns and intensity of surface precipitation at high spatial resolution over stratiform and convective precipitation regimes. The evaluations reveal the potential of extracting relevant, high spatial features over complex surface types such as mountainous regions and coastlines. The algorithm allows users to interpret the inputs’ importance and can serve as a starting point for further exploration of precipitation systems within the field of hydrometeorology.

Study of Ecosystem Degradation Dynamics in the Peruvian Highlands: Landsat Time-Series Trend Analysis (1985–2022) with ARVI for Different Vegetation Cover Types

The high-Andean vegetation ecosystems of the Bombón Plateau in Peru face increasing degradation due to aggressive anthropogenic land use and the climate change scenario. The lack of historical degradation evolution information makes implementing adaptive monitoring plans in these vulnerable ecosystems difficult. Remote sensor technology emerges as a fundamental resource to fill this gap. The objective of this article was to analyze the degradation of vegetation in the Bombón Plateau over almost four decades (1985–2022), using high spatiotemporal resolution data from the Landsat 5, 7, and 8 sensors. The methodology considers: (i) the use of the atmosphere resistant vegetation index (ARVI), (ii) the implementation of non-parametric Mann–Kendall trend analysis per pixel, and (iii) the affected vegetation covers were determined by supervised classification. This article’s results show that approximately 13.4% of the total vegetation cover was degraded. According to vegetation cover types, bulrush was degraded by 21%, tall grass by 18%, cattails by 16%, wetlands by 14%, and puna grass by 13%. The Spearman correlation (p &lt; 0.01) determined that degraded covers are replaced by puna grass and change factors linked with human activities. Finally, this article concludes that part of the vegetation degradation is related to anthropogenic activities such as agriculture, overgrazing, urbanization, and mining. However, the possibility that environmental factors have influenced these events is recognized.

Effect of spatial resolution of climatological data on streamflow simulations using the SWAT : A case study

Data quality always affects the accuracy of model output. Rainfall is the basic data required in hydrological modelling as rainfall to runoff conversion is the core of all such models. Regional modelling studies required high resolution spatio-temporal data and availability of data at appropriate resolution also greatly affect the modelling results. Therefore, efforts have been started to record climatic variables at finer resolution so that they will be useful for block level and gram Panchayat level studies. In this study, an effort has been made to identify the effect of using various resolution climatic data on streamflow simulation in the Kesinga catchment of the Mahanadi river basin. Three types of rainfall sets with spatial resolution of 0.25° × 0.25° and 1° × 1° from IMD and one set of recorded rainfall data of the Special Relief Commissioner (SRC), Govt. of Odisha is used in combination with IMD 1° × 1° gridded temperature to simulate streamflow at the Kesinga gauging station using the Soil and Water Assessment Tool (SWAT) keeping other parameters constant. The three simulations were analyzed using NSE, R2, RMSE, PBIAS, P-factor and R-factor. The results depicted that IMD gridded rainfall data sets predicted similar flows compared to the SRC recorded rainfall data which proves the fairness of IMD gridded data is at par with the recorded rainfall data of SRC, Govt. of Odisha.

Leveraging high spatiotemporal resolution data of pesticides applied to agricultural fields in California to identify toxicity reduction opportunities

Pesticides remain a leading environmental hazard, imperiling aquatic and terrestrial ecosystems. Reducing pesticide toxicity is hampered by the ability to evaluate toxicity over large extents, the spatiotemporal resolution of pesticide use data, the ability to assess cumulative toxicity, and the identification of health/economic contributions of different pesticide application sites. We introduce the Environmental Release Tool, a sub-tool of the Pesticide Mitigation Prioritization Model, to advance these four areas. Using daily pesticide use reports required for agricultural applicators in California, we quantify the applied toxicity of pesticides to fish as well as aquatic invertebrates, nonvascular plants, and vascular plants. With the tool’s ability to quantify applied toxicity for hundreds of pesticides and watersheds simultaneously, we explore the significance of accounting for cumulative applied pesticide toxicity for application sites and watersheds statewide. Our results show that 14 pesticides account for 99.9% of applied toxicity, and 16 of 432 application site types introduce 90% of toxicity for taxa investigated. We also find cumulative applied toxicity within watersheds was significantly greater (p &lt;1.0 E-16) than the maximum impact pesticide for all taxonomic groups, with a mean-annual difference of 460–630%. While cumulative applied toxicity was significant, and sources varied in individual watersheds, the net applied toxicity can be approximated with a short list of active ingredients and site types.

Estimating fishing effort in small-scale fisheries using high-resolution spatio-temporal tracking data (an implementation framework illustrated with case studies from Portugal)

Small-scale fisheries (SSF, boats < 12 m) represent 90% of this sector at a worldwide scale and 84% of the EU fleet. Mapping the areas and intensity where the fishing operations occur is essential for spatial planning, safety, fisheries sustainability and biodiversity conservation. The EU is currently regulating position tracking of SSF fishing vessels requiring precision resolved geo-positional data (sec to min resolution).Here we developed a series of procedures aimed at categorizing fishing boats behaviour using high resolution data. Our integrated approach involve novel routines aimed at (i) produce an expert validated data set, (ii) pre-processing of positional data, (iii) establishing minimal required temporal resolution, and (iv) final assessment of an optimized classification model. Objective (iv) was implemented by using statistical and machine learning (ML) routines, using novel combinations of fixed thresholds estimates using regression trees and classification methods based on anti-mode, Gaussian Mixture Models (GMM), Expectation Maximisation (EM) algorithms, Hidden Markov Models (HMM) and Random Forest (RF). Of relevance, the final evaluation framework incorporates both error quantification and fishing effort indicators. We tested the method by running through four SSF fisheries from Portugal recorded every 30 sec, with 183 boat trips validated, and concluded that the more robust time interval for data acquisition in these metiers should be <2 min and that mode and random forest methods with pre-data treatment gave the best results. A special effort was concentrated in a visual support provided by the results produced by this new method, making its interpretation easier, thus facilitating transference and translation into other fishery levels. After the current validation in the Portuguese SSF fleet, we posit that our novel procedure has the potential to serve as an integrated quantitative approach to the EU SSF management.

Bayesian estimation of real-time epidemic growth rates using Gaussian processes: local dynamics of SARS-CoV-2 in England

Abstract Quantitative assessments of the recent state of an epidemic and short-term projections for the near future are key public-health tools that have substantial policy impacts, helping to determine if existing control measures are sufficient or need to be strengthened. Key to these quantitative assessments is the ability to rapidly and robustly measure the speed with which an epidemic is growing or decaying. Frequently, epidemiological trends are addressed in terms of the (time-varying) reproductive number R. Here, we take a more parsimonious approach and calculate the exponential growth rate, r, using a Bayesian hierarchical model to fit a Gaussian process to the epidemiological data. We show how the method can be employed when only case data from positive tests are available, and the improvement gained by including the total number of tests as a measure of the heterogeneous testing effort. Although the methods are generic, we apply them to SARS-CoV-2 cases and testing in England, making use of the available high-resolution spatio-temporal data to determine long-term patterns of national growth, highlight regional growth, and spatial heterogeneity.

The use of crowdsourced social media data to improve flood forecasting

Reliable flood forecasting systems are essential for predicting and mitigating the impact of flooding worldwide. However, minimising flood forecast uncertainties remains a challenging task due to many sources of uncertainty in underlying flood simulation modelling. Such uncertainties can be reduced by employing data assimilation techniques to dynamically incorporate the most recent available observations into the system while accounting for existing uncertainties in both models and observations. However, traditional observations often lack the necessary temporal or spatial resolution, limiting the adoption of data assimilation methods for real-time applications. In contrast, data collection through crowdsourcing has grown in popularity with the potential to provide high spatiotemporal resolution data, especially in urban areas. Nevertheless, the use of crowdsourcing is still impacted by validation uncertainties and data quality, which makes it a complementing data to traditional observations rather than an alternative data source. This paper presents a novel methodology for assimilating crowdsourced social media data to improve a 2D flood forecasting model through various update strategies. The methodology was tested against a real case flood event of the 2017 Phetchaburi flood (Thailand), and the performance of different update strategies was evaluated with reference to the calibrated model output obtained from a particle swarm optimisation algorithm. Empirical results demonstrate that global state updates suffer from inconsistencies in predicted water levels, whereas topographically based local state updates provide encouraging results. Specifically, the improvement due to the local state update alone is short-lived, and findings indicate that a longer lasting improvement in flood forecasting performance can be achieved through a combination of both state and boundary updates. Overall, the results indicate the feasibility of utilising crowdsourced social media data to improve the performance of flood forecasting systems for urban environments.