Prediction for Various Drought Classes Using Spatiotemporal Categorical Sequences

Drought frequently spreads across large spatial and time scales and is more complicated than other natural disasters that can damage economic and other natural resources worldwide. However, improved drought monitoring and forecasting techniques can help to minimize the vulnerability of society to drought and its consequent influences. This emphasizes the need for improved drought monitoring tools and assessment techniques that provide information more precisely about drought occurrences. Therefore, this study developed a new method, Model-Based Clustering for Spatio-Temporal Categorical Sequences (MBCSTCS), that uses state selection procedures through finite mixture modeling and model-based clustering. The MBCSTCS uses the functional structure of first-order Markov model components for modeling each data group. In MBCSTCS, the suitable order K of the components is selected by Bayesian information criterion (BIC). In MBCSTCS, the estimated mixing proportions and the posterior probabilities are used to compute probability distribution associated with the future steps of transitions. Furthermore, MBCSTCS predicts drought occurrences in future time using spatiotemporal categorical sequences of various drought classes. The MBCSTCS is applied to the six meteorological stations in the northern area of Pakistan. Moreover, it is found that MBCSTCS provides expeditious information for the long-term spatiotemporal categorical sequences. These findings may be helpful to make plans for early warning systems, water resource management, and drought mitigation policies to decrease the severe effects of drought.

Download Full-text

A Parallel EM Algorithm for Model-Based Clustering Applied to the Exploration of Large Spatio-Temporal Data

Technometrics ◽

10.1080/00401706.2013.826146 ◽

2013 ◽

Vol 55 (4) ◽

pp. 513-523 ◽

Cited By ~ 5

Author(s):

Wei-chen Chen ◽

George Ostrouchov ◽

David Pugmire ◽

Prabhat ◽

Michael Wehner

Keyword(s):

Em Algorithm ◽

Temporal Data ◽

Model Based Clustering ◽

Model Based ◽

Spatio Temporal

Download Full-text

Building an Improved Drought Climatology Using Updated Drought Tools: A New Mexico Food-Energy-Water (FEW) Systems Focus

Frontiers in Climate ◽

10.3389/fclim.2020.576653 ◽

2020 ◽

Vol 2 ◽

Author(s):

Lindsay E. Johnson ◽

Hatim M. E. Geli ◽

Michael J. Hayes ◽

Kelly Helm Smith

Keyword(s):

New Mexico ◽

The United States ◽

Decision Makers ◽

Drought Monitoring ◽

Drought Risk ◽

Human Environment ◽

Food Energy ◽

Management Actions ◽

Monitoring Tools ◽

Effects Of Drought

Drought is a familiar climatic phenomenon in the United States Southwest, with complex human-environment interactions that extend beyond just the physical drought events. Due to continued climate variability and change, droughts are expected to become more frequent and/or severe in the future. Decision-makers are charged with mitigating and adapting to these more extreme conditions and to do that they need to understand the specific impacts drought has on regional and local scales, and how these impacts compare to historical conditions. Tremendous progress in drought monitoring strategies has occurred over the past several decades, with more tools providing greater spatial and temporal resolutions for a variety of variables, including drought impacts. Many of these updated tools can be used to develop improved drought climatologies for decision-makers to use in their drought risk management actions. In support of a Food-Energy-Water (FEW) systems study for New Mexico, this article explores the use of updated drought monitoring tools to analyze data and develop a more holistic drought climatology applicable for New Mexico. Based upon the drought climatology, droughts appear to be occurring with greater frequency and magnitude over the last two decades. This improved drought climatology information, using New Mexico as the example, increases the understanding of the effects of drought on the FEW systems, allowing for better management of current and future drought events and associated impacts.

Download Full-text

Clustering the Concentrations of PM10 and O3: Application of Spatio-temporal Model-based Clustering

10.21203/rs.3.rs-264277/v1 ◽

2021 ◽

Author(s):

parisa saeipourdizaj ◽

saeed musavi ◽

Akbar Gholampour ◽

parvin sarbakhsh

Keyword(s):

Air Pollution ◽

Mixture Model ◽

Average Concentration ◽

Optimal Number ◽

Number Of Clusters ◽

Model Based Clustering ◽

Model Based ◽

Spatio Temporal ◽

Temporal Dimensions ◽

Optimal Number Of Clusters

Abstract Air pollution data are large-scale dataset which can be analyzed in low scales by clustering to recognize the pattern of pollution and have simpler and more comprehensible interpretation. So, this study aims to cluster the days of year 2017 according to the hourly O3 and PM10 amounts collected from four stations of Tabriz by using spatio-temporal mixture model-based clustering (STMC). Besides, mixture model-based clustering with temporal dimension (TMC) and mixture model-based clustering without considering spatio-temporal dimensions (MC) were utilized to compare with STMC. To evaluate the efficiency of these three models and obtain the optimal number of clusters in each model, BIC and ICL criteria were used. According to BIC and ICL, STMC outperforms TMC and MC. Three clusters for O3 and four clusters for PM10 were selected as the optimal number of clusters to fit STMC models. Regarding PM10, the average concentration was the highest in cluster 4. Regarding O3, all summer days were in cluster 3 and the average concentration of this cluster was the highest. Cluster 2 had the lowest concentration with a high difference from clusters 1 and 3 and its average temperature was the lowest. Autumn days make up about 84% of this cluster. The clustering of polluted and clean days into separate groups and observing the effect of meteorological factors on the amount of concentration in each cluster clearly prove the efficiency of the model. Results of STMC showed that efficiency of clustering in air pollution data increases by considering both spatio-temporal dimensions.

Download Full-text

Dynamic model-based clustering for spatio-temporal data

Statistics and Computing ◽

10.1007/s11222-017-9735-9 ◽

2017 ◽

Vol 28 (2) ◽

pp. 359-374 ◽

Cited By ~ 3

Author(s):

Lucia Paci ◽

Francesco Finazzi

Keyword(s):

Dynamic Model ◽

Temporal Data ◽

Model Based Clustering ◽

Model Based ◽

Spatio Temporal

Download Full-text

Model-Based Clustering and Classification for Data Science

10.1017/9781108644181 ◽

2019 ◽

Cited By ~ 17

Author(s):

Charles Bouveyron ◽

Gilles Celeux ◽

T. Brendan Murphy ◽

Adrian E. Raftery

Keyword(s):

Data Science ◽

Model Based Clustering ◽

Model Based ◽

Clustering And Classification

Download Full-text

Lane Detection Model Based on Spatio-Temporal Network With Double Convolutional Gated Recurrent Units

IEEE Transactions on Intelligent Transportation Systems ◽

10.1109/tits.2021.3060258 ◽

2021 ◽

pp. 1-13

Author(s):

Jiyong Zhang ◽

Tao Deng ◽

Fei Yan ◽

Wenbo Liu

Keyword(s):

Lane Detection ◽

Temporal Network ◽

Detection Model ◽

Model Based ◽

Spatio Temporal ◽

Gated Recurrent Units

Download Full-text

Influence of Random Forest Hyperparameterization on Short-Term Runoff Forecasting in an Andean Mountain Catchment

Atmosphere ◽

10.3390/atmos12020238 ◽

2021 ◽

Vol 12 (2) ◽

pp. 238

Author(s):

Pablo Contreras ◽

Johanna Orellana-Alvear ◽

Paul Muñoz ◽

Jörg Bendix ◽

Rolando Célleri

Keyword(s):

Random Forest ◽

Model Performance ◽

Early Warning Systems ◽

Point Of View ◽

Lead Times ◽

Physical Parameters ◽

Runoff Forecasting ◽

Spatio Temporal ◽

Improved Model ◽

Search Approach

The Random Forest (RF) algorithm, a decision-tree-based technique, has become a promising approach for applications addressing runoff forecasting in remote areas. This machine learning approach can overcome the limitations of scarce spatio-temporal data and physical parameters needed for process-based hydrological models. However, the influence of RF hyperparameters is still uncertain and needs to be explored. Therefore, the aim of this study is to analyze the sensitivity of RF runoff forecasting models of varying lead time to the hyperparameters of the algorithm. For this, models were trained by using (a) default and (b) extensive hyperparameter combinations through a grid-search approach that allow reaching the optimal set. Model performances were assessed based on the R2, %Bias, and RMSE metrics. We found that: (i) The most influencing hyperparameter is the number of trees in the forest, however the combination of the depth of the tree and the number of features hyperparameters produced the highest variability-instability on the models. (ii) Hyperparameter optimization significantly improved model performance for higher lead times (12- and 24-h). For instance, the performance of the 12-h forecasting model under default RF hyperparameters improved to R2 = 0.41 after optimization (gain of 0.17). However, for short lead times (4-h) there was no significant model improvement (0.69 < R2 < 0.70). (iii) There is a range of values for each hyperparameter in which the performance of the model is not significantly affected but remains close to the optimal. Thus, a compromise between hyperparameter interactions (i.e., their values) can produce similar high model performances. Model improvements after optimization can be explained from a hydrological point of view, the generalization ability for lead times larger than the concentration time of the catchment tend to rely more on hyperparameterization than in what they can learn from the input data. This insight can help in the development of operational early warning systems.

Download Full-text