Spatiotemporal heterogeneity in meteorological and hydrological drought patterns and propagations influenced by climatic variability, LULC change, and human regulations

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: March 12, 2024

Abstract This study aims to quantify meteorological–hydrological drought propagations and examine the potential impacts by climatic variability, LULC change (LULC), human regulations. An integrated observation-modeling framework quantifies propagation intervals assesses mechanisms influencing hydrological droughts. Meteorological droughts are characterized using Standardized Precipitation Evapotranspiration Index (SPEI), assessed through Streamflow (SSI) across diverse zones. Cross-correlation analysis between SPEI SSI time series identifies lag associated with highest correlation as interval. Mechanisms investigated via a coupled empirical-process modeling incorporating Soil Water Assessment Tool (SWAT). Discrepancies simulated observed help extent of regulation on characteristics propagation. The Yellow River Basin (YRB), divided into six subzones based climate characteristics, is selected case study. Key findings include: (1) were extremely severe most YRB during 1990s, while 2000s showed some mitigation primarily due precipitation increases. (2) Hydrological times from meteorology hydrology demonstrated substantial spatiotemporal variability. In general, summer shorter than other seasons. (3) Propagation in arid regions cropland or built-up land cover versus grassland woodland, reverse held for humid regions. (4) Human regulations prolonged times, likely reservoir designed overcome water deficits. While focus this paper, methodologies applicable worldwide enhance forecasting resource management. various contexts worldwide.

Language: Английский

---

Enhancing Drought Forecast Accuracy Through Informer Model Optimization

Land, Journal Year: 2025, Volume and Issue: 14(1), P. 126 - 126

Published: Jan. 9, 2025

As droughts become more frequent due to climate change and shifts in land use, enhancing the accuracy of drought prediction is becoming crucial for informed water resource management. This study employed Informer model forecast conducted a comparative analysis with Autoregressive Integrated Moving Average (ARIMA), long short-term memory (LSTM), Convolutional Neural Network (CNN) models. The findings indicate that outperforms other three models terms forecasting across all time scales. Nevertheless, predictive capacity remains suboptimal when it comes intervals. Aiming at problem short scale, this proposed named VMD-JAYA-Informer based on Variational Mode Decomposition (VMD) JAVA optimization algorithm improve model. VMD-JAYA-ARIMA, VMD-JAYA-LSTM, VMD-JAYA-CNN, performance these was evaluated using root mean square error (RMSE), Nash–Sutcliffe efficiency coefficient (NSE), Mean Absolute Error (MAE). model’s 1-month SPEI significantly surpasses alternative demonstrates robust agreement actual data. Simultaneously, exhibits equally optimal different In order validate model, four meteorological stations Songliao River Basin were chosen random. validation results demonstrate scale (NSE values 0.8663, 0.8765, 0.8822, 0.8416, respectively). Additionally, scales, further demonstrating its generalizability excellence shorter

Language: Английский

---

Global horizontal irradiance prediction model for multi-site fusion under different aerosol types

Renewable Energy, Journal Year: 2024, Volume and Issue: 227, P. 120565 - 120565

Published: April 25, 2024

Language: Английский

---

Deep learning model for the deformation prediction of concrete dams under multistep and multifeature inputs based on an improved autoformer

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 137, P. 109109 - 109109

Published: Aug. 12, 2024

Language: Английский

---

Multi-step ahead forecasting of daily streamflow based on the transform-based deep learning model under different scenarios

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 14, 2025

Predicting runoff with precision holds immense importance for flood control, water resource management, and basin ecological dispatch. Deep learning, especially long short-term memory (LSTM) neural networks, has excelled in prediction, often outperforming traditional hydrological models. Recent studies suggest that deep learning models employing the self-attention mechanism, such as Transformer Informer, can achieve even better results than LSTM. However, research exploring multi-step prediction capabilities of these novel across diverse scenarios remains scarce. In this investigation, we introduce a relative location coding-enhanced Informer model, termed Rel-Informer, compare its performance rainfall-runoff against standard Transformer, LSTM The publicly available CAMELS dataset is utilized training validating models, four experiments are designed: (1) Individual modeling (one model per catchment); (2) Regional region); (3) Fine-tuned regional (fine-tuned from Experiment 2); (4) Large-scale ungauged catchments all catchments). findings reveal Rel-Informer consistently performs other particularly predictions (1–3 days ahead). Although less precise individual modeling, it significantly benefits fine-tuning. large-scale effectively predicts catchments, showcasing potential widespread prediction. This study underscores influence characteristics, snowmelt baseflow indices, on accuracy. conclusion, enhanced improved position encoding, emerges promising tool forecasting, data-rich catchments.

Language: Английский

---

Enhancement of standardized precipitation evapotranspiration index predictions by machine learning based on regression and soft computing for Iran’s arid and hyper–arid region

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(3), P. e0319678 - e0319678

Published: March 18, 2025

Drought is a climate risk that affects access to safe water, crop development, ecological stability, and food production. Therefore, developing drought prediction methods can lead better management of surface groundwater resources. Similarly, machine learning be used find improved relationships between nonlinear variables in complex systems. Initially, the standardized precipitation evapotranspiration index (SPEI) was calculated, then using large–scale signals such as (the North Atlantic Oscillation, Arctic Pacific Decadal Southern Oscillation Index), along with climatic including temperature, precipitation, potential evapotranspiration, predictions were made for period 1966–2014. Several new models Least Square Support Vector Regression (LSSVR), Group Method Data Handling (GMDH), Multivariate Adaptive Splines (MARS) prediction. The results showed estimating SPEI moderately arid climates, GMDH model criteria (RMSE = 0.26, MAE 0.17, NSE 0.95 validation) under scenario S1 (included all plus previous month) performed better, while cold LSSVR 0.22, 0.18, S1, hot climate, 0.29, 0.19, 0.93 S2 meteorological had higher accuracy. Although MARS less accurate validation, it accuracy during calibration compared other two climates. predicting beneficial. It concluded are useful tools different climates within similar ranges.

Language: Английский

---

Research on speed optimization of fixed route ship with low data dependence

Ocean Engineering, Journal Year: 2025, Volume and Issue: 328, P. 121065 - 121065

Published: March 26, 2025

Language: Английский

---

A Daily Reference Crop Evapotranspiration Forecasting Model Based on Improved Informer

Agriculture, Journal Year: 2025, Volume and Issue: 15(9), P. 933 - 933

Published: April 25, 2025

Daily reference crop evapotranspiration (ET0) is crucial for precision irrigation management, yet traditional prediction methods struggle to capture its dynamic variations due the complexity and nonlinearity of meteorological conditions. To address this, we propose an Improved Informer model enhance ET0 accuracy, providing a scientific basis agricultural water management. Using soil data from Yingde region, employed Maximal Information Coefficient (MIC) identify key influencing factors integrated Residual Cycle Forecasting (RCF), Star Aggregate Redistribute (STAR), Fully Adaptive Normalization (FAN) techniques into model. MIC analysis identified total shortwave radiation, sunshine duration, maximum temperature at 2 m, 28–100 cm depth, surface pressure as optimal features. Under five-feature scenario (S3), improved achieved superior performance compared Long Short-Term Memory (LSTM) original models, with MAE reduced 0.065 (LSTM: 0.637, Informer: 0.171) MSE 0.007 0.678, 0.060). The inference time was also by 31%, highlighting enhanced computational efficiency. effectively captures periodic nonlinear characteristics ET0, offering novel solution management significant practical implications.

Language: Английский

---

Investigating the Performance of the Informer Model for Streamflow Forecasting

Water, Journal Year: 2024, Volume and Issue: 16(20), P. 2882 - 2882

Published: Oct. 10, 2024

Recent studies have shown the potential of transformer-based neural networks in increasing prediction capacity. However, classical transformers present several problems such as computational time complexity and high memory requirements, which make Long Sequence Time-Series Forecasting (LSTF) challenging. The contribution to series flood events using deep learning techniques is examined, with a particular focus on evaluating performance Informer model (a implementation transformer architecture), attempts address previous issues. predictive capabilities are explored compared statistical methods, stochastic models traditional networks. accuracy, efficiency well limits approaches demonstrated via numerical benchmarks relating real river streamflow applications. Using daily flow data from River Test England main case study, we conduct rigorous evaluation efficacy capturing complex temporal dependencies inherent series. analysis extended encompass diverse datasets various locations (>100) United Kingdom, providing insights into generalizability Informer. results highlight superiority over established forecasting especially regarding LSTF problem. For forecast horizon 168 days, achieves an NSE 0.8 maintains MAPE below 10%, while second-best (LSTM) only −0.63 25%, respectively. Furthermore, it observed that dependence structure series, expressed by climacogram, affects network.

Language: Английский

---

CNN-Informer: A Hybrid Deep Learning Model for Seizure Detection on Long-term EEG

Neural Networks, Journal Year: 2024, Volume and Issue: 181, P. 106855 - 106855

Published: Oct. 28, 2024

Language: Английский

---