Bayesian Spatio-temporal Model Research Articles

A Novel Bayesian Spatio-Temporal Surveillance Metric to Predict Emerging Infectious Disease Areas of High Disease Risk.

Identification of areas of high disease risk has been one of the top goals for infectious disease public health surveillance. Accurate prediction of these regions leads to effective resource allocation and faster intervention. This paper proposes a novel prediction surveillance metric based on a Bayesian spatio-temporal model for infectious disease outbreaks. Exceedance probability, which has been commonly used for cluster detection in statistical epidemiology, was extended to predict areas of high risk. The proposed metric consists of three components: the area's risk profile, temporal risk trend, and spatial neighborhood influence. We also introduce a weighting scheme to balance these three components, which accommodates the characteristics of the infectious disease outbreak, spatial properties, and disease trends. Thorough simulation studies were conducted to identify the optimal weighting scheme and evaluate the performance of the proposed prediction surveillance metric. Results indicate that the area's own risk and the neighborhood influence play an important role in making a highly sensitive metric, and the risk trend term is important for the specificity and accuracy of prediction. The proposed prediction metric was applied to the COVID-19 case data of South Carolina from March 12, 2020, and the subsequent 30 weeks of data.

Effects of climate variability on the spatio-temporal distribution of Dengue in Valle del Cauca, Colombia, from 2001 to 2019.

Dengue is a vector-borne disease that has increased over the past two decades, becoming a global public health emergency. The transmission of dengue is contingent upon various factors, among which climate variability plays a significant role. However, there remains substantial uncertainty regarding the underlying mechanisms. This study aims to investigate the spatial and temporal patterns of dengue risk and to quantify the associated risk factors in Valle del Cauca, Colombia, from 2001 to 2019. To achieve this, a spatio-temporal Bayesian hierarchical model was developed, integrating delayed and non-linear effects of climate variables, socio-economic factors, along with spatio-temporal random effects to account for unexplained variability. The results indicate that average temperature is positively associated with dengue risk 0-2 months later, showing a 35% increase in the risk. Similarly, high precipitation levels lead to increased risk approximately 2-3 months later, while relative humidity showed a constant risk within a 6 months-lag. These findings could be valuable for local health authorities interested in developing early warning systems to predict future risks in advance.

Gradient boosting: A computationally efficient alternative to Markov chain Monte Carlo sampling for fitting large Bayesian spatio-temporal binomial regression models

Disease forecasting and surveillance often involve fitting models to a tremendous volume of historical testing data collected over space and time. Bayesian spatio-temporal regression models fit with Markov chain Monte Carlo (MCMC) methods are commonly used for such data. When the spatio-temporal support of the model is large, implementing an MCMC algorithm becomes a significant computational burden. This research proposes a computationally efficient gradient boosting algorithm for fitting a Bayesian spatio-temporal mixed effects binomial regression model. We demonstrate our method on a disease forecasting model and compare it to a computationally optimized MCMC approach. Both methods are used to produce monthly forecasts for Lyme disease, anaplasmosis, ehrlichiosis, and heartworm disease in domestic dogs for the contiguous United States. The data have a spatial support of 3108 counties and a temporal support of 108–138 months with 71–135 million test results. The proposed estimation approach is several orders of magnitude faster than the optimized MCMC algorithm, with a similar mean absolute prediction error.

Bayesian spatio-temporal modeling of the Brazilian fire spots between 2011 and 2022

Wildfires are among the most common natural disasters in many world regions and actively impact life quality. These events have become frequent due to climate change, other local policies, and human behavior. Fire spots are areas where the temperature is significantly higher than in the surrounding areas and are often used to identify wildfires. This study considers the historical data with the geographical locations of all the “fire spots” detected by the reference satellites covering the Brazilian territory between January 2011 and December 2022, comprising more than 2.2 million fire spots. This data was modeled with a spatio-temporal generalized linear mixed model for areal unit data, whose inferences about its parameters are made in a Bayesian framework and use meteorological variables (precipitation, air temperature, humidity, and wind speed) and a human variable (land-use transition and occupation) as covariates. The meteorological variables humidity and air temperature showed the most significant impact on the number of fire spots for each of the six Brazilian biomes.

Bayesian spatio-temporal analysis of dengue transmission in Lao PDR

Dengue, a zoonotic viral disease transmitted by Aedes mosquitoes, poses a significant public health concern throughout the Lao People’s Democratic Republic (Lao PDR). This study aimed to describe spatial–temporal patterns and quantify the effects of environmental and climate variables on dengue transmission at the district level. The dengue data from 2015 to 2020 across 148 districts of Lao PDR were obtained from the Lao PDR National Center for Laboratory and Epidemiology (NCLE). The association between monthly dengue occurrences and environmental and climate variations was investigated using a multivariable Zero-inflated Poisson regression model developed in a Bayesian framework. The study analyzed a total of 72,471 dengue cases with an incidence rate of 174 per 100,000 population. Each year, incidence peaked from June to September and a large spike was observed in 2019. The Bayesian spatio-temporal model revealed a 9.1% decrease (95% credible interval [CrI] 8.9%, 9.2%) in dengue incidence for a 0.1 unit increase in monthly normalized difference vegetation index at a 1-month lag and a 5.7% decrease (95% CrI 5.3%, 6.2%) for a 1 cm increase in monthly precipitation at a 6-month lag. Conversely, dengue incidence increased by 43% (95% CrI 41%, 45%) for a 1 °C increase in monthly mean temperature at a 3-month lag. After accounting for covariates, the most significant high-risk spatial clusters were detected in the southern regions of Lao PDR. Probability analysis highlighted elevated trends in 45 districts, emphasizing the importance of targeted control strategies in high-risk areas. This research underscores the impact of climate and environmental factors on dengue transmission, emphasizing the need for proactive public health interventions tailored to specific contexts in Lao PDR.

Interactions of multiple stressors on the Bombay-duck Harpadon nehereus population in a complex estuarine ecosystem

In an era marked by unprecedented anthropogenic change, marine systems are increasingly subjected to interconnected and dynamic external stressors, which profoundly reshape the behavior and resilience of marine ecological components. Nevertheless, despite widespread recognition of the significance of stressor interactions, there persist notable knowledge deficits in quantifying their interactions and the specific biological consequences that result. To bridge this crucial gap, this research detected and examined the causal relationships between five key exogenous stressors in a complex estuarine ecosystem. Furthermore, a Bayesian Hierarchical Spatio-temporal modeling framework was proposed to quantitatively evaluate the distinct, interactive, and globally sensitive effects of multiple stressors on the population dynamics of a crucial fish species: Harpadon nehereus. The results showed that interactions were detected between fisheries pressure (FP), the Pacific Decadal Oscillation index (PDO), runoff volume (RV), and sediment load (SL), with five of these interactions producing significant synergistic effects on H. nehereus biomass. The SL*PDO and RV*PDO interactions had positive synergistic effects, albeit through differing processes. The former interaction amplified the individual effects of each stressor, while the latter reversed the direction of the original impact. Indeed overall, the synergistic effect of multiple stressors was not favorable, with FP in particular posing the greatest threat to H. nehereus population. This threat was more pronounced at high SL or negative PDO phases. Therefore, local management efforts aimed at addressing multiple stressors and protecting resources should consider the findings. Additionally, although the velocity of climate change (VoCC) failed to produce significant interactions, changes in this stressor had the most sensitive impacts on the response of H. nehereus population. This research strives to enhance the dimensionality, generalizability, and flexibility of the quantification framework for marine multi-stressor interactions, aiming to foster broader research collaboration and jointly tackle the intricate pressures facing marine ecosystems.

Spatiotemporal patterns and socioeconomic determinants of pulmonary tuberculosis in Dongguan city, China, during 2011–2020: an ecological study

ObjectivePulmonary tuberculosis (PTB) is a critical challenge worldwide, particularly in China. This study aimed to explore the spatiotemporal transmission patterns and socioeconomic factors of PTB in Dongguan city, China.Methods/designAn ecological...

Bayesian spatio-temporal modeling to assess the effect of land-use changes on the incidence of Cutaneous Leishmaniasis in the Brazilian Amazon

Cutaneous Leishmaniasis (CL) is a vector-borne disease caused by a protozoan of the genus Leishmania and is considered one of the most important neglected tropical diseases. The Brazilian Amazon Forest harbors one of the highest diversity of Leishmania parasites and vectors and is one of the main focuses of the disease in the Americas. Previous studies showed that some types of anthropogenic disturbances have affected the abundance and distribution of CL vectors and hosts; however, few studies have thoroughly investigated the influence of different classes of land cover and land-use changes on the disease transmission risk. Here, we quantify the effect of land use and land-cover changes on the incidence of CL in all municipalities within the Brazilian Amazon Forest, from 2001 to 2017. We used a structured spatiotemporal Bayesian model to assess the effect of forest cover, agriculture, livestock, extractivism, and- deforestation on CL incidence, accounting for confounding variables such as population, climate, socioeconomic, and spatiotemporal random effects. We found that the increased risk of CL was associated with deforestation, especially modulated by a positive interaction between forest cover and livestock. Landscapes with ongoing deforestation for extensive cattle ranching are typically found in municipalities within the Amazon Frontier, where a high relative risk for CL was also identified. These findings provide valuable insights into developing effective public health policies and land-use planning to ensure healthier landscapes for people.

Applying a Bayesian Spatiotemporal Model to Examine Excess County-Level Cardiovascular Disease Death Rates During the COVID-19 Pandemic.

Amid the COVID-19 pandemic, national cardiovascular disease (CVD) death rates increased, especially among younger adults. County-level variation has not been documented. Using county-level CVD deaths (ICD-10 codes: I00-I99) from the US National Vital Statistics System, we developed a Bayesian multivariate spatiotemporal model to estimate excess CVD death rates in 2020 based on trends from 2010-2019 for adults aged 35-64 and ≥65 years. Among adults aged 35-64 years, 64.7% of counties experienced significant excess CVD death rates. The median county-level CVD death rate in 2020 was 150 per 100,000 persons, which exceeded the predicted rate for 2020 (median excess death rate: 11 per 100,000; median excess rate ratio: 1.08). Among adults aged ≥65 years, 15.2% of counties experienced significant excess CVD death rates. The median county-level CVD death rate was 1,546 per 100,000 in 2020, which exceeded the predicted rate in 2020 (median excess death rate: 48 per 100,000, median excess rate ratio: 1.03). Counties with significant excess death rates in 2020 were geographically dispersed. In 2020, disruptions of county-level CVD death rates were widespread, especially among younger adults, suggesting the continued importance of CVD prevention and treatment in younger adults in communities across the country.

Bayesian spatio-temporal modelling of environmental, climatic, and socio-economic influences on malaria in Central Vietnam

BackgroundDespite the successful efforts in controlling malaria in Vietnam, the disease remains a significant health concern, particularly in Central Vietnam. This study aimed to assess correlations between environmental, climatic, and socio-economic factors in the district with malaria cases.MethodsThe study was conducted in 15 provinces in Central Vietnam from January 2018 to December 2022. Monthly malaria cases were obtained from the Institute of Malariology, Parasitology, and Entomology Quy Nhon, Vietnam. Environmental, climatic, and socio-economic data were retrieved using a Google Earth Engine script. A multivariable Zero-inflated Poisson regression was undertaken using a Bayesian framework with spatial and spatiotemporal random effects with a conditional autoregressive prior structure. The posterior random effects were estimated using Bayesian Markov Chain Monte Carlo simulation with Gibbs sampling.ResultsThere was a total of 5,985 Plasmodium falciparum and 2,623 Plasmodium vivax cases during the study period. Plasmodium falciparum risk increased by five times (95% credible interval [CrI] 4.37, 6.74) for each 1-unit increase of normalized difference vegetation index (NDVI) without lag and by 8% (95% CrI 7%, 9%) for every 1ºC increase in maximum temperature (TMAX) at a 6-month lag. While a decrease in risk of 1% (95% CrI 0%, 1%) for a 1 mm increase in precipitation with a 6-month lag was observed. A 1-unit increase in NDVI at a 1-month lag was associated with a four-fold increase (95% CrI 2.95, 4.90) in risk of P. vivax. In addition, the risk increased by 6% (95% CrI 5%, 7%) and 3% (95% CrI 1%, 5%) for each 1ºC increase in land surface temperature during daytime with a 6-month lag and TMAX at a 4-month lag, respectively. Spatial analysis showed a higher mean malaria risk of both species in the Central Highlands and southeast parts of Central Vietnam and a lower risk in the northern and north-western areas.ConclusionIdentification of environmental, climatic, and socio-economic risk factors and spatial malaria clusters are crucial for designing adaptive strategies to maximize the impact of limited public health resources toward eliminating malaria in Vietnam.

Estimation of excess mortality attributable to extreme heat during the summer of 2022 in Catalonia, Spain. Bayesian spatio-temporal model

Estimation of excess mortality attributable to extreme heat during the summer of 2022 in Catalonia, Spain. Bayesian spatio-temporal model

Temperature, relative humidity and elderly type 2 diabetes mortality: A spatiotemporal analysis in Shandong, China

BackgroundThe mortality of type 2 diabetes mellitus (T2DM) can be affected by environmental factors. However, few studies have explored the effects of environmental factors across diverse regions over time. Given the vulnerability observed in the elderly group in previous research, this research applied Bayesian spatiotemporal models to assess the associations in the elderly group. MethodsData on T2DM death in the elderly group (aged over 60 years old) at the county level were collected from the National Death Surveillance System between 1st January 2013 and 31st December 2019 in Shandong Province, China. A Bayesian spatiotemporal model was employed with the integrated Nested Laplace Approach to explore the associations between socio-environmental factors (i.e., temperatures, relative humidity, the Normalized Difference Vegetation Index (NDVI), particulate matter with a diameter of 2.5 μm or less (PM2.5) and gross domestic product (GDP)) and T2DM mortality. ResultsT2DM mortality in the elderly group was found to be associated with temperature and relative humidity (i.e., temperature: Relative Risk (RR) = 1.41, 95% Credible Interval (CI): 1.27–1.56; relative humidity: RR = 1.05, 95% CI:1.03–1.06), while no significant associations were found with NDVI, PM2.5 and GDP. In winter, significant impacts from temperature (RR = 1.18, 95% CI: 1.06–1.32) and relative humidity (RR = 0.94, 95% CI: 0.89–0.99) were found. Structured and unstructured spatial effects, temporal trends and space-time interactions were considered in the model. ConclusionsHigher mean temperatures and relative humidities increased the risk of elderly T2DM mortality in Shandong Province. However, a higher humidity level decreased the T2DM mortality risk in winter in Shandong Province. This research indicated that the spatiotemporal method could be a useful tool to assess the impact of socio-environmental factors on health by combining the spatial and temporal effects.

Factors associated with the spatiotemporal distribution of dog rabies in Tunisia.

Despite continuous efforts of veterinary services to control rabies in dogs since 1982, rabies remains a cause of death in Tunisia, with more than five reported human cases in 2022. As little is known on the determinants of transmission of rabies in dogs, better understand which factors contribute to its spatial heterogeneity in Tunisia is critical for developing bespoke mitigation activities. In this context, we developed Bayesian Poisson mixed-effect spatio-temporal model upon all cases of rabid dogs reported in each delegation during the period from 2019 to 2021. The best fitting model highlighted the association between the risk of rabies and the mean average monthly temperature, the density of markets and the density of dogs in delegations. Interestingly, no relationship was found between intensity of vaccination in dogs and the risk of rabies. Our results provided insights into the spatio-temporal dynamics of dog rabies transmission and highlighted specific geographic locations where the risk of infection was high despite correction for associated explanatory variables. Such an improved understanding represent key information to design bespoke, cost-efficient, rabies prevention and control strategies to support veterinary services activities and policymaking.

Association between air pollution, meteorological factors and hand, foot, and mouth disease incidence in Zhejiang Province: A Bayesian spatio-temporal modelling study

Temperature and humidity are known to influence seasonal patterns of hand, foot and mouth disease (HFMD) transmission. However, the effect of air pollution and climate-air pollution interaction on the intensity of HFMD epidemics is poorly understood. This study aims to quantify the exposure-response effects of air pollution on HFMD risk by level of meteorological indicators in Zhejiang Province. A spatiotemporal Bayesian hierarchical model was used to determine the exposure-response association between air pollution, meteorological factors and the relative risk (RR) of HFMD. The model was fitted to monthly HFMD case data for the 92 counties in Zhejiang Province between January 2013, and December 2021. We assessed both linear interaction effects between air pollutants and meteorological factors on HFMD, and the modification role of extreme weather conditions on the nonlinear air pollutant-HFMD association. A 10 μg/m³ increase in PM10 and NO2 concentration raised the HFMD risk by 6.6% (RR = 1.066, 95% CI: 1.042, 1.092) and 20.8% (RR = 1.208, 95% CI: 1.162, 1.256) respectively. HFMD was negatively associated with O3 (RR = 0.980, 95% CI: 0.968, 0.993), SO2 (RR = 0.820, 95% CI: 0.776, 0.868) and CO (RR = 0.770, 95% CI: 0.621, 0.956). Temperature, humidity, and precipitation positively interacted with most air pollutants on HFMD, while wind speed had a negative interaction effect. In higher temperature conditions, the HFMD risk associated with PM10 increased. Elevated humidity levels amplified the risk associated with PM10, as did CO concentrations within a certain range. Heavy rainfall intensified the risk of HFMD associated with PM10, SO2, and CO. Results of this study showed that air pollution was associated with HFMD incidence, with different climate factors playing various modification roles in the associations and primarily leading to increased risk under extreme conditions. These findings provide a theoretical basis for personal protection and government prevention and control measures.

The hospitalisation risk of chronic circulatory and respiratory diseases associated with coal mining in the general population in Queensland, Australia

Queensland is the main coal mining state in Australia where populations in coal mining areas have been historically exposed to coal mining emissions. Although a higher risk of chronic circulatory and respiratory diseases has been associated with coal mining globally, few studies have investigated these associations in the Queensland general population. This study estimates the association of coal production with hospitalisations for chronic circulatory and respiratory diseases in Queensland considering spatial and temporal variations during 1997–2014.An ecological analysis used a Bayesian hierarchical spatiotemporal model to estimate the association of coal production with standardised rates of each, chronic circulatory and respiratory diseases, adjusting for sociodemographic factors and considering the spatial structure of Queensland's statistical areas (SA2) in the 18-year period. Two specifications; with and without a space-time interaction effect were compared using the integrated nested Laplace approximation -INLA approach. The posterior mean of the best fit model was used to map the spatial, temporal and spatiotemporal trends of risk.The analysis considered 2,831,121 hospitalisation records. Coal mining was associated with a 4 % (2.4—5.5) higher risk of hospitalisation for chronic respiratory diseases in the model with a space-time interaction effect which had the best fit. An emerging higher risk of either chronic circulatory and respiratory diseases was identified in eastern areas and some coal-mining areas in central and southeast Queensland. There were important disparities in the spatiotemporal trend of risk between coal -and non-coal mining areas for each, chronic circulatory and respiratory diseases.Coal mining is associated with an increased risk of chronic respiratory diseases in the Queensland general population. Bayesian spatiotemporal analyses are robust methods to identify environmental determinants of morbidity in exposed populations. This methodology helps identifying at-risk populations which can be useful to support decision-making in health. Future research is required to investigate the causality links between coal mining and these diseases.

The role of prescription opioid and cannabis supply policies on opioid overdose deaths.

Mandatory prescription drug monitoring programs and cannabis legalization have been hypothesized to reduce overdose deaths. We examined associations between prescription monitoring programs with access mandates ("must-query PDMPs"), legalization of medical and recreational cannabis supply, and opioid overdose deaths in United States counties in 2013-2020. Using data on overdose deaths from the National Vital Statistics System, we fit Bayesian spatiotemporal models to estimate risk differences and 95% credible intervals (CrI) in county-level opioid overdose deaths associated with enactment of these state policies. Must-query PDMPs were independently associated with on average 0.8 (95% CrI: 0.5, 1.0) additional opioid-involved overdose deaths per 100,000 person-years. Legal cannabis supply was not independently associated with opioid overdose deaths in this time period. Must-query PDMPs enacted in the presence of legal (medical or recreational) cannabis supply were associated with 0.7 (95% CrI: 0.4, 0.9) more opioid-involved deaths, relative to must-query PDMPs without any legal cannabis supply. In a time when overdoses are driven mostly by non-prescribed opioids, stricter opioid prescribing policies and more expansive cannabis legalization were not associated with reduced overdose death rates.

Spatio-temporal forecasting for the US Drought Monitor

Abstract The US Drought Monitor is the leading drought monitoring tool in the United States. Updated weekly and freely distributed, it records the drought conditions as geo-referenced polygons showing one of six ordered levels. These levels are determined by a mixture of quantitative environmental measurements and local expert opinion across the entire United States. At present, forecasts of the Drought Monitor only convey the expected direction of drought development (i.e. worsen, persist, subside) and do not communicate any uncertainty. This limits the utility of forecasts. In this paper, we describe a Bayesian spatio-temporal ordinal hierarchical model for use in modelling and projecting drought conditions. The model is flexible, scalable, and interpretable. By viewing drought data as areal rather than point-referenced, we reduce the cost of sampling from the posterior by avoiding dense matrix inversion. Draws from the posterior predictive distribution produce future forecasts of actual drought levels—rather than only the direction of drought development—and all sources of uncertainty are propagated into the posterior. Spatial random effects and an autoregressive model structure capture spatial and temporal dependence, and help ensure smoothness in forecasts over space and time. The result is a framework for modelling and forecasting drought levels and capturing forecast uncertainty.

Invited commentary: Motivating better methods-and better data collection-for measuring prevalence of drug misuse.

The United States continues to suffer a drug overdose crisis that has resulted in over 100,000 deaths annually since 2021. Despite decades of attention, estimates of the prevalence of drug use at the spatiotemporal resolutions necessary for resource allocation and intervention evaluation are lacking. Current approaches to measure prevalence of drug use, such as population surveys, capture-recapture, and multiplier methods, have significant limitations. Santaella-Tenorio et al. (Am J Epidemiol. XXXX;XXX(XX):XXXX-XXXX)) use a novel joint Bayesian spatiotemporal modeling approach to estimate county-level opioid misuse prevalence in New York state from 2007 to 2018 and identify significant intra-state variation. By leveraging five data sources and simultaneously modeling different opioid-related outcomes - such as deaths, emergency department visits, and treatment visits - they obtain policy-relevant insights into the prevalence of opioid misuse and opioid-related outcomes at high spatiotemporal resolutions. This study provides future researchers with a sophisticated modeling approach that allows them to incorporate multiple data sources in a rigorous statistical framework. The limitations of the study reflect the constraints of the broader field and underscores the importance of enhancing current surveillance with better, newer, and more timely data that is both standardized and easily accessible to inform public health policies and interventions.

Comparing lagged impacts of mobility changes and environmental factors on COVID-19 waves in rural and urban India: a Bayesian spatiotemporal modelling study.

Previous research in India has identified urbanisation, human mobility and population demographics as key variables associated with higher district level COVID-19 incidence. However, the spatiotemporal dynamics of mobility patterns in rural and urban areas in India, in conjunction with other drivers of COVID-19 transmission, have not been fully investigated. We explored travel networks within India during two pandemic waves using aggregated and anonymized weekly human movement datasets obtained from Google, and quantified changes in mobility before and during the pandemic compared with the mean baseline mobility for the 8-week time period at the beginning of 2020. We fit Bayesian spatiotemporal hierarchical models coupled with distributed lag non-linear models (DLNM) within the integrated nested Laplace approximate (INLA) package in R to examine the lag-response associations of drivers of COVID-19 transmission in urban, suburban, and rural districts in India during two pandemic waves in 2020-2021. Model results demonstrate that recovery of mobility to 99% that of pre-pandemic levels was associated with an increase in relative risk of COVID-19 transmission during the Delta wave of transmission. This increased mobility, coupled with reduced stringency in public intervention policy and the emergence of the Delta variant, were the main contributors to the high COVID-19 transmission peak in India in April 2021. During both pandemic waves in India, reduction in human mobility, higher stringency of interventions, and climate factors (temperature and precipitation) had 2-week lag-response impacts on the R t of COVID-19 transmission, with variations in drivers of COVID-19 transmission observed across urban, rural and suburban areas. With the increased likelihood of emergent novel infections and disease outbreaks under a changing global climate, providing a framework for understanding the lagged impact of spatiotemporal drivers of infection transmission will be crucial for informing interventions.

Spatio-Temporal Fluctuation Analysis of Ecosystem Service Values in Northeast China over Long Time Series: Based on Bayesian Hierarchical Modeling

Ecosystems are undergoing continuous degradation due to the dual perturbation of global climate change and human activities, posing unprecedented threats and challenges to the ecosystem services they provide. To gain a deeper understanding of the spatio-temporal evolution of ecosystem service value (ESV), it is essential to accurately capture the characteristics of its spatial and temporal changes and its influencing factors. However, traditional spatio-temporal statistical methods are limited to analyzing the heterogeneity of ESV in a single temporal or spatial dimension, which fails to meet the comprehensive analysis needs for spatio-temporal heterogeneity over an extended continuum. Therefore, this paper constructs a Bayesian spatio-temporal hierarchical model to analyze the ESV heterogeneity in both temporal and spatial dimensions in Northeast China from 2000 to 2020 to accurately identify the regions with unstable fluctuations in ESV and analyze the influencing factors behind them. It aims to comprehensively and systematically reveal the intrinsic laws of spatio-temporal evolution of ESV, and provide a scientific basis for relevant decision-making. The study found a continuous fluctuating downward trend of ESV in Northeast China from 2000 to 2020, with significant spatial and temporal heterogeneity. Notably, the distribution of hot and cold spots is regularly concentrated, especially in the transition zone from low hills to plains, which forms an “unstable zone” of spatial and temporal fluctuations of ESV. Natural factors such as NDVI and NPP exhibit a significant positive correlation with ESV, while social factors like population density and GDP show a strong negative correlation. Compared to traditional statistical methods, the Bayesian spatio-temporal hierarchical model, with its outstanding flexibility and accuracy, provides a new perspective and way of thinking for analyzing classical spatio-temporal problems. Firstly, the model examines time and space as a whole and fully accounts for the influence of spatio-temporal interactions on ESV. Secondly, the Bayesian spatio-temporal hierarchical model meets the needs of long-term continuous ESV outcome detection, which provides us with solid support for a deeper understanding of the evolution of ESV.