Meteorological Stations Research Articles

Long-Term Wind and Air Temperature Patterns in the Southeastern Region of Iran through Model Simulation and Ground Observations

Dust storms are one of the important natural hazards that affect the lives of inhabitants all around the world, especially in North Africa and the Middle East. In this study, wind speed, wind direction, and air temperature patterns are investigated in one of the dustiest cities in Sistan Basin, Zahedan City, located in southeast Iran, over a 17-year period (2004–2020) using a WRF model and ground observation data. The city is located near a dust source and is mostly affected by local dust storms. The World Meteorology Organization (WMO) dust-related codes show that the city was affected by local dust, with 52 percent of the total dust events occurring during the period (2004–2021). The city’s weather station reported that 17.5% and 43% were the minimum and maximum dusty days, respectively, during 2004–2021. The summer and July were considered the dustiest season and month in the city. Since air temperature, wind speed, and wind direction are important factors in dust rising and propagation, these meteorological factors were simulated using the Weather Research and Forecasting (WRF) model for the Zahedan weather station. The WRF model’s output was found to be highly correlated with the station data; however, the WRF simulation mostly overestimated when compared with station data during the study period (2004–2020). The model had a reasonable performance in wind class frequency distribution at the station, demonstrating that 42.6% of the wind was between 0.5 and 2, which is in good agreement with the station data (42% in the range of 0.5–2). So, the WRF model effectively simulated the wind class frequency distribution and the wind direction at Zahedan station, despite overestimating the wind speed as well as minimum, maximum, and average air temperatures during the 17-year period.

The Evaluation of Rainfall Forecasting in a Global Navigation Satellite System-Assisted Numerical Weather Prediction Model

Accurate water vapor information is crucial for improving the quality of numerical weather forecasting. Previous studies have incorporated tropospheric water vapor data obtained from a global navigation satellite system (GNSS) into numerical weather models to enhance the accuracy and reliability of rainfall forecasts. However, research on evaluating forecast accuracy for different rainfall levels and the development of corresponding forecasting platforms is lacking. This study develops and establishes a rainfall forecasting platform supported by the GNSS-assisted weather research and forecasting (WRF) model, quantitatively assessing the effect of GNSS precipitable water vapor (PWV) on the accuracy of WRF model forecasts for light rain (LR), moderate rain (MR), heavy rain (HR), and torrential rain (TR). Three schemes are designed and tested using data from seven ground meteorological stations in Xi’an City, China, in 2021. The results show that assimilating GNSS PWV significantly improves the forecast accuracy of the WRF model for different rainfall levels, with the root mean square error (RMSE) improvement rates of 8%, 15%, 19%, and 25% for LR, MR, HR, and TR, respectively. Additionally, the RMSE of rainfall forecasts demonstrates a decreasing trend with increasing magnitudes of assimilated PWV, particularly effective in the range of [50, 55) mm where the lowest RMSE is 3.58 mm. Moreover, GNSS-assisted numerical weather model shows improvements in statistical forecasting indexes such as probability of detection (POD), false alarm rate (FAR), threat score (TS), and equitable threat score (ETS) across all rainfall intensities, with notable improvements in the forecasts of HR and TR. These results confirm the high precision, visualization capabilities, and robustness of the developed rainfall forecasting platform.

Use of LCZs with urban canopy modeling to evaluate urban growth effects on meteorological conditions in the Salt Lake Valley

Urban Canopy Models (UCMs) can be used to represent urban characteristics on a mesoscale without accounting for effects of individual buildings. A WRF-based single-layer urban canopy model (SLUCM) was validated for the greater Salt Lake City area (GSLA) using measurements from two weather stations for a one-week period during summer 2019. A baseline model based on current GSLA properties developed using the Noah land surface model, SLUCM, and refined urban classifications based on WUDAPT Local Climate Zones (LCZs) produced acceptable predictions and computational times. The baseline model was validated using measured meteorological data and clearly showed diurnal heating patterns. Three scenarios representing multi-decade growth were defined by modifying LCZs in different parts of the GSLA. Compared to baseline values, results showed increased urban density increased local nighttime temperatures 5–10 °C depending on the extent of growth, but did not affect daytime temperatures. This was due to changes in LCZ anthropogenic heating rates and sky view factors which were most noticeable during early nighttime hours, consistent with atmospheric heat island behavior. Predicted 2-m relative humidity and 10-m wind speeds changed minimally.

Characteristics of Sand-Driving Wind Regime and Sand Drift Potential in Sandy Areas on Both Sides of Longyangxia Reservoir in China

This study focuses on the desertification areas of Talatan and Mugetan, located on both sides of the Longyangxia Reservoir in Qinghai Province, China. These areas pose serious threats to the safety of the reservoir due to frequent wind-sand activities. The research aims to provide a scientific basis for effective sand control by deeply analyzing the dynamic change of wind.The wind speed and direction data from Shazhuyu and Guinan meteorological stations in Qinghai Province, China were used to calculate the numerical and directional characteristics of sand-driving wind and sand drift potential.The results indicate that: (1) The number of sand-driving wind at Talatan is about 7 times that at Mugetan, and the maximum and average wind speeds are about 7 m/s larger than those at Mugetan. At Talatan, the primary sand-driving wind direction is exclusively West Northwest (WNW). Conversely, at Mugetan, the sand-driving wind direction remains westward in spring, autumn, and winter, shifting eastward solely in summer. (2) The annual sand drift potential of Talatan and Mugetan is less than 200VU, two sandy areas are both in low wind energy environments; (3) The major causes of the significant differences in sand-driving wind regimes between the Talatan and Mugetan are topography, precipitation, wind direction stability.These conclusions offer valuable insights for developing targeted sand control strategies, thereby mitigating the potential hazards posed by wind-sand activities to the Longyangxia Reservoir and its surrounding ecosystems.

The coupling of winds, ocean turbulence, and High Salinity Shelf Water in the Terra Nova Bay Polynya

The Terra Nova Bay (TNB) Polynya in the Western Ross Sea of Antarctica is a major producer of High Salinity Shelf Water (HSSW), a precursor to Antarctica Bottom Water (AABW). Processes occurring in and around the polynya can therefore effect change in the lower limb of overturning circulation in this region. Here, we use data from a densely-instrumented upper-ocean mooring, deployed for 1 year in a region of active HSSW formation within TNB, to examine the coupling of surface brine rejection and vertical mixing to katabatic wind forcing. We find a high correlation between salinity and winds during the wintertime HSSW production season at the mooring site, with a lag-response of 20 h in near-surface (∼47 m) salinity to winds measured at the nearby Automatic Weather Station (AWS) Manuela. Salinity and temperature measurements show a fully destratified water column by June, with a lag-response of near-seabed (∼360 m) salinity to near-surface salinity of just 5 h. Measurements of turbulent kinetic energy (TKE) dissipation rate (ɛ) from moored pulse-coherent acoustic Doppler current profilers (ADCPs) show general agreement with classic boundary layer scaling (BLS), and calculations of a vertical mixing timescale using the Obukhov length scale average to ∼2.5 h during austral winter, consistent with the 5-h lag time observed in the salinity data. Comparisons to data from concurrent mooring deployments along the southern boundary of TNB, as well as to previously published assessments of model simulations and data from Climatic Long-term Interaction for the Mass-balance in Antarctica (CLIMA) moorings, allow us to explore spatial variability in the coupling of winds and salinity across TNB and to speculate on possible HSSW circulation pathways.

Unveiling urban heat: harnessing personal weather stations for enhanced daily mapping of heat stress across North Carolina

Unveiling urban heat: harnessing personal weather stations for enhanced daily mapping of heat stress across North Carolina

Predicting Energy Production in Renewable Energy Power Plants Using Deep Learning

It is very important to analyze and forecast energy production for investments in renewable energy resources. In this study, the energy production of wind and solar power plants, which are among the leading renewable energy sources, was estimated using deep learning. For a solar power plant, three different solar power plants with 1MW installed power were examined. Three-year energy production data of power plants were taken. These data were used with the deep learning method long short-term memory (LSTM) and seasonal autoregressive moving average (SARIMA). Results were obtained for each dataset; they were subjected to five different (MSE, RMSE, NMSE, MAE, and MAPE) error performance measurement systems. In the LSTM model, the highest accuracy rate was 81% and the lowest accuracy rate was 59%. In the SARIMA model, the highest accuracy rate was 66% and the lowest accuracy rate was 41%. As for wind energy, wind speeds in two different places were estimated. Wind speed data were taken from meteorological stations. Datasets were tested with MAPE, R2, and RMSE error performance measurement systems. LSTM, GRU, CNN-LSTM, CNN-RNN, LSTM-GRU, and CNN-GRU deep learning methods were used in this study. The CNN-GRU model achieved a maximum accuracy of 99.81% in wind energy forecasting.

Unveiling the persistence of meteorological drought in Iraq: a comprehensive spatiotemporal analysis

Droughts, stemming from prolonged periods of scarce precipitation, lead to significant economic and societal consequences. Iraq has recurrently experienced the impact of severe droughts. This study assessed meteorological drought over Iraq, employing the Standardized Precipitation Index (SPI) on 12-month time scales from 2000 to 2022. The SPI values were calculated for 27 meteorological stations over Iraq based on the monthly datasets. The analysis of the SPI highlighted that the years 2000, 2008, 2009, 2010, 2012, 2018, and 2022 stood out as the driest in Iraq from 2000 to 2022. Notably, 2022 emerged as a particularly severe drought year, with the Ramadi weather station recording the maximum SPI value at -3.48. Different time scales in the analysis provided a more precise depiction of drought severity and frequency in Iraq. Namely, the findings resulting from this study offer a nuanced understanding of the magnitude, intensity, frequency, and duration of drought events in Iraq. This nuanced approach equips stakeholders with valuable insights for implementing more cost-effective mitigation and adaptation measures. This study contributes crucial information for understanding and addressing droughts challenges in Iraq. The comprehensive analysis aids in formulating targeted strategies to alleviate the impact of droughts on people and ecosystems.

An Evaluation of a Winter Mortality Model for the Mountain Pine Beetle

The winter mortality of mountain pine beetle (Dendroctonus ponderosae Hopkins) larvae caused by temperatures below survivable thresholds and sudden decreases in temperature is among the most influential factors limiting population growth. Due to the importance of winter mortality in the demise of mountain pine beetle infestations, a widely used winter mortality model was developed at the Canadian Forest Service and the United States Forest Service. It predicts lethal temperatures and survival probabilities given temperature time series over the winter season. We present a rare and possibly the first peer-reviewed validation of this winter mortality model, wherein we independently tested the model at a new region in Canada by comparing model predictions to the observed lower lethal temperature thresholds and cold-associated mortality. Model predictions were biologically reasonable but slightly biased. Bias was exacerbated by the inaccurate translation of air temperature data from weather stations to temperatures under the bark where larvae develop. The spatial prediction of relative mortality observed across the study area in Banff National Park was poor—likely because the mountainous terrain presents a difficult prediction challenge when under-bark temperatures are not directly observed. Our results will help inform users of model constraints and how to optimize the accuracy of model predictions.

Design and experiment of a lightweight cloud particle imager

Abstract Accurate in-situ measurement of cloud microphysical parameters such as particle diameter, number concentration and droplet spectrum distribution are of great significance in cloud physics, climate change, numerical weather forecasting and weather modification. This work describes the design and characterization of a newly developed in situ measurement instrument for cloud and fog named lightweight cloud particle imager (LCPI). The basic measurement principle of LCPI is based on scattering light imaging of particles in a dark field. A ring LED lighting source was designed to improve the image quality. Field measurements were carried out at a high-altitude research station on Lushan Mountain. The measured size distributions showed good agreement with parallel measurements of a cloud droplet spectrometer (fog monitor-120). The field data also showed that LCPI was able to measure the size distribution of cloud droplets from 5 to 300 μm with a high spatial resolution. In addition, another observation experiment carried out in the Tuli River weather station showed that LCPI also had the ability to acquire clear images of ice crystal particles larger than 10 μm. Thus, LCPI provides an opportunity to simultaneously quantify the microphysical structure of different cloud types.

Structure and Trends in Climate Parameters of Wine-Growing Regions in Slovenia

This study examined the structure and trends of climate parameters important for grape production from 1952 to 2022 in the wine-growing regions of Podravje, Posavje, and Primorska in Slovenia. Average and extreme temperature and precipitation data from six meteorological stations in three wine-growing regions were divided into annual and growing seasons. The results show that in the period 1991–2022, there was a warming in the growing season in all regions by 1.4–1.7 °C, except the southern part of Primorska (Koper station) 0.6 °C, compared to the reference period 1961–1990. The heat accumulation indices (GDDs and HI) increased significantly, which is mainly due to the increase in the maximum temperature in the growing season temperature (GST max) and the number of days with Tmax > 30 °C (NDT30). The NDT30 increased the most, by a factor of more than four. In the reference period (1961–1990), however, the trend in the number of hot days was even slightly negative. The mean seasonal temperature rose to around 17 °C in regions with a continental climate and to around 19 °C in the Mediterranean region, which could be reflected in the earlier ripening of the grapes. The trends show a decrease in total annual precipitation (AP) after 1991, but this was significant only at one inland location (Maribor), while the total precipitation during the growing season (GSP) decreased significantly at three locations (Maribor, Bilje, and Koper).

High-Resolution WRF Modeling of Wind and Thermal Regimes with LCZ in Almaty, Kazakhstan

This study evaluates the effectiveness of the Weather Research and Forecasting (WRF) model in simulating high-resolution atmospheric conditions for Almaty, Kazakhstan, a city prone to stagnant winter air. While the previously used Bougeault and Lacarrere scheme for parameterizing the planetary boundary layer was applied in high-resolution modeling, the number of vertical levels was increased, and a detailed local climate zones (LCZs) map was included. Ground-based observations from meteorological stations and monitoring stations, remote sensing data, and radiosonde measurements are used to verify the model. Comparison results with ground-based observations show that the WRF model with the LCZ map provides a better representation of the wind and thermal regimes of Almaty compared to the three-class land use map, including in high resolution. A good correspondence of wind direction is demonstrated by comparing the modeling results with pollutant transport plumes recorded by remote sensing data. In addition, a good correlation was found between land surface temperature from satellite data and air temperature simulated by WRF with a resolution of 333 m. A comparison of simulated data and aerological measurements confirmed that downscaling did not have a significant impact on boundary layer calculations. Analysis of turbulent processes showed that the adopted model effectively describes the attenuation and dissipation of turbulent kinetic energy and reflects the typical diurnal variations of meteorological processes in the atmosphere of Almaty in the anticyclonic winter period. The results of high-resolution WRF modeling can form the basis for the development of a hybrid system capable of modeling atmospheric processes at the building level.

Drivers of long-term changes in summer compound hot extremes in China: Climate change, urbanization, and vegetation greening

Induced by numerous environmental and anthropogenic factors, the frequency of compound hot extremes (CoHot) is increasing globally. However, there is a lack of a comprehensive understanding of the relative importance of different factors in the change of CoHot, as most studies focused solely on individual factors. We systematically attribute the long-term trends of summertime CoHot in China to 24 factors including climate background, climate change, urbanization (including impervious area expansion and three-dimensional urban morphology), and vegetation change, using partial least square regression (PLSR) and variable importance projections for analysis. Cross-evaluation results based on 2259 meteorological stations indicate that the PLSR model accurately simulates the trends of CoHot frequency, duration, and intensity, with a relative error of 4–20% from national to individual station scales. A warm and humid climate background, rising average summertime temperatures, impervious area expansions, high building volume densities, low building heights, and vegetation wilting are the main factors that favoring the increasing and intensifying trends of CoHot. Although summertime warming contributes the most, the significance of urbanization and vegetation greening has been growing in recent years. Urbanization's contribution increases from 9.3% on a national scale to over 40% in urban centers, mainly due to increased building volume density rather than expansions of impervious area. In northern China, vegetation greening has reduced the local increasing trends in frequency/duration of CoHot by 20% to 40%, and such mitigation effects are primarily induced by increased vegetation cover. Our results suggest that, reducing building volume density and promoting vegetation greening can mitigate the increasing risk of CoHot under global warming.

An Approach for Future Droughts in Northwest Türkiye: SPI and LSTM Methods

Predetermining the risk of possible future droughts enables proactive measures to be taken in key areas such as agriculture, water management, and food security. Through these predictions, governments, non-governmental organizations, and farmers can develop water-saving strategies, encourage more efficient use of water, and minimize economic losses that may occur due to drought. Thus, future drought forecasts stand out as a strategic planning tool for the protection of natural resources. To achieve this aim, forecasted drought conditions for the next decade (2024–2034) at nine meteorological stations in the Sakarya basin, located northwest of Türkiye, are examined, using historical monthly precipitation data from 1991 to 2023. This study uses the Standardized Precipitation Index (SPI) and Long Short-Term Memory (LSTM) deep learning methods to investigate future meteorological droughts. The research confirms the compatibility and reliability of the LSTM method for forecasting meteorological droughts by comparing historical and forecasted SPI values’ correlograms and trends. In addition, drought maps are created to visually represent the spatial distribution of the most severe droughts expected in the coming years, and areas at risk of drought in the Sakarya Basin are determined. The study contributes to the limited literature on forward-looking drought forecasts in the Sakarya Basin and provides valuable information for long-term water resource planning and drought management in the region.

Development of a stochastic rainfall generator to yield unprecedented rainfall events

Unprecedented rainfall events are defined as rainfall events that have extremely high magnitude with low probability of occurrence. Although such events have occurred in the recent past, there might be no historical record of unprecedented rainfall. While significant efforts have been made to understand the major drivers of unprecedented events, the use of statistical modelling to estimate the likelihood of unprecedented events has gained little to no attention. The current study proposes a ‘serial’ type stochastic rainfall generator (SRG) to simulate daily rainfall with an additional focus on unprecedented events. The study introduces resampling of model parameters to yield the non-stationarity in simulated rainfall distribution. The proposed approach is applied to simulate daily rainfall over twenty-six grid points in Southeast Texas, which experienced Hurricane Harvey in 2017. Towards this, the study obtains rainfall accumulation data during Harvey from eight weather stations during. We found that the proposed approach can successfully generate unprecedented events if two power law tail tuning parameters are adjusted with respect to future warming levels. Rainfall magnitudes correspond to 50, 100, 250 and 500-year return periods simulated by the proposed model are found to be substantially higher than historical (i.e., observed) return period. Furthermore, the study found that the estimated return periods of Hurricane Harvey like rainfall is almost the same as the return period estimated by a former study of Emmanuel (2017). The potential benefit of the proposed approach in hydrological risk assessment has also been discussed.

Operational and experimental snow observation systems in the upper Rofental: data from 2017 to 2023

Abstract. This publication presents a comprehensive hydrometeorological data set for three research sites in the upper Rofental (1891–3772 m a.s.l., Ötztal Alps, Austria) and is a companion publication to a data collection published in 2018. The time series presented here comprise data from 2017 to 2023 and originate from three meteorological and snow hydrological stations at 2737, 2805, and 2919 m a.s.l. The fully equipped automatic weather stations include a specific set of sensors to continuously record snow cover properties. These are automatic measurements of snow depth, snow water equivalent, volumetric solid and liquid water contents, snow density, layered snow temperature profiles, and snow surface temperature. One station is extended by a particular arrangement of two snow depth and water equivalent recording devices to observe and quantify wind-driven snow transport. These devices are installed at nearby wind-exposed and sheltered locations and are complemented by an acoustic-based snow drift sensor. We present data for temperature, precipitation, humidity, wind speed, and radiation fluxes and explore the continuous snow measurements by combined analyses of meteorological and snow data to show typical seasonal snow cover characteristics. The potential of the snow drift observations is demonstrated with examples of measured wind speeds, snow drift rates, and redistributed snow amounts during several blowing snow events. The data complement the scientific monitoring infrastructure in the research catchment and represent a unique time series of high-altitude mountain weather and snow observations. They enable comprehensive insights into the dynamics of high-altitude meteorological and snow processes and are collected to support the scientific community, local stakeholders, and the interested public, as well as operational warning and forecasting services. The data are publicly available from the GFZ Data Services repository: https://doi.org/10.5880/fidgeo.2023.037 (Department of Geography, University of Innsbruck, 2024).

Development of a Daily Cloud-Free Snow-Cover Dataset Using MODIS-Based Snow-Cover Probability for High Mountain Asia during 2000–2020

Investigating the changes in snow cover caused by climate change is extremely important and has attracted increasing attention in cryosphere and climate research. Optimal remote sensing-based snow datasets can provide long-term daily and global spatial-temporal snow-cover distribution at regional and global scales. However, the application of these snow-cover products is inevitably limited because of the space–time discontinuities caused by cloud obscuration, which poses a significant challenge in snowpack-related studies, especially in High Mountain Asia (HMA), an area that has high-elevation mountains, complex terrain, and harsh environments and has fewer observation stations. To address this issue, we developed an improved five-step hybrid cloud removal strategy by integrating the daily merged snow-cover probability (SCP) algorithm, eight-day merged SCP algorithm, decision tree algorithm, temporal downscaling algorithm, and optimal threshold segmentation algorithm to produce a 21-year, daily cloud-free snow-cover dataset using two daily MODIS snow-cover products over the HMA. The accuracy assessment demonstrated that the newly developed cloud-free snow-cover product achieved a mean overall accuracy of 93.80%, based on daily classified snow depth observations from 86 meteorological stations over 10 years. The time series of the daily percentage of binary snow-cover over HMA was analyzed during this period, indicating that the maximum snow cover tended to change more dramatically than the minimum snow cover. The annual snow-cover duration (SCD) experienced an insignificantly increasing trend over most of the northeastern and southwestern HMA (e.g., Qilian, eastern Kun Lun, the east of Inner Tibet, the western Himalayas, the central Himalayas, and the Hindu Kush) and an insignificant declining trend over most of the northwestern and southeastern HMA (e.g., the eastern Himalayas, Hengduan, the west of Inner Tibet, Pamir, Hissar Alay, and Tien). This new high-quality snow-cover dataset will promote studies on climate systems, hydrological modeling, and water resource management in this remote and cold region.

A Hierarchical Model to Predict Time of Flowering of Kiwifruit Using Weather Data and Budbreak Dynamics.

Accurate prediction of flowering times is essential for efficient orchard management for kiwifruit, facilitating timely pest and disease control and pollination interventions. In this study, we developed a predictive model for flowering time using weather data and observations of budbreak dynamics for the 'Hayward' and 'Zesy002' kiwifruit. We used historic data of untreated plants collected from 32 previous studies conducted between 2007 and 2022 and analyzed budbreak and flowering timing alongside cumulative heat sum (growing degree days, GDDs), chilling unit (CU) accumulation, and other environmental variables using weather data from the weather stations nearest to the study orchards. We trained/parameterized the model with data from 2007 to 2019, and then evaluated the model's efficacy using testing data from 2020 to 2022. Regression models identified a hierarchical structure with the accumulation of GDDs at the start of budbreak, one of the key predictors of flowering time. The findings suggest that integrating climatic data with phenological events such as budbreak can enhance the predictability of flowering in kiwifruit vines, offering a valuable tool for kiwifruit orchard management.

Performance Evaluation of Artificial Neural Networks for Estimating Refer-ence Evapotranspiration in Shahat, Libya using limited climatic data

This study was conducted with the aim of evaluating the performance of artificial neural networks (ANNs) to estimate the reference evapotranspiration using limited climate data in Shahat region in Libya, compared to using the FAO Penman-Monteith equation (FPM), which requires temperature, wind speed, relative humidity and number of sunshine hours, which are rarely available in most meteorological stations in developing countries. In this study, we used the average temperature (Tmean) and the average relative humidity (RHmean) obtained from Shahat meteorological station for the period from 1963 to 1999, and the extraterrestrial radiation (Ra), which can be calculated given the location and time of the day. These data are divided into two groups, from 1963 to 1988 and from 1989 to 1999 for the training and validation phases of the neural networks, respectively. This study concluded that using (Tmean), (RHmean) and (Ra) gave the best agreement with the results calculated with the FAO Penman-Monteith equation, where the values of R2 and RMSE are equal to 0.98 and 0.26, respectively.

Prediction of DEDI index for meteorological drought with the VMD-CBiLSTM hybrid model

Drought can cause great harm and loss, so accurate and efficient drought prediction has certain research significance. In previous studies, a single machine learning model is often used to predict the factors related to drought, such as precipitation and temperature, to indirectly explain the influence degree of drought, but the overall prediction accuracy is low, which can not fully and effectively predict the nonlinear and non-stationary drought characteristic information. In this paper, the variational modal decomposition model (VMD) is used to decompose the meteorological drought time series, and the convolution neural network (CNN) and the improved bidirectional long short-term memory neural network (BiLSTM) are combined to construct the meteorological drought hybrid prediction model (VMD-CBiLSTM). The research results show that using VMD-CBiLSTM model to forecast the monthly daily evapotranspiration deficit index (DEDI) of three weather stations, compared with the results predicted by VMD-LSTM, VMD-BiLSTM and VMD-ARIMA models, the average prediction accuracy R2 is increased by 0.076, 0.034 and 0.328 respectively, and the average RMSE is decreased by 0.178, 0.094 and 1.373 respectively. Compared with single model, VMD-CBiLSTM can not only reduce the uncertainty of meteorological drought prediction caused by climate model heterogeneity, but also improve the accuracy of drought prediction in arid and semi-arid regions, which can provide reference for coping with drought occurrence and drought early warning in advance.