Temperature Forecasts Research Articles

Recurrent neural network modeling of multivariate time series and its application in temperature forecasting.

Temperature forecasting plays an important role in human production and operational activities. Traditional temperature forecasting mainly relies on numerical forecasting models to operate, which takes a long time and has higher requirements for the computing power and storage capacity of computers. In order to reduce computation time and improve forecast accuracy, deep learning-based temperature forecasting has received more and more attention. Based on the atmospheric temperature, dew point temperature, relative humidity, air pressure, and cumulative wind speed data of five cities in China from 2010 to 2015 in the UCI database, multivariate time series atmospheric temperature forecast models based on recurrent neural networks (RNN) are established. Firstly, the temperature forecast modeling of five cities in China is established by RNN for five different model configurations; secondly, the neural network training process is controlled by using the Ridge Regularizer (L2) to avoid overfitting and underfitting; and finally, the Bayesian optimization method is used to adjust the hyper-parameters such as network nodes, regularization parameters, and batch size to obtain better model performance. The experimental results show that the atmospheric temperature prediction error based on LSTM RNN obtained a minimum error compared to using the base models, and these five models obtained are the best models for atmospheric temperature prediction in the corresponding cities. In addition, the feature selection method is applied to the established models, resulting in simplified models with higher prediction accuracy.

Intelligent Fuzzy Models: WM, ANFIS, and Patch Learning for the Competitive Forecasting of Environmental Variables

This paper focuses on the application of fuzzy modeling methods in the field of environmental engineering. Since predicting meteorological data is considered to be a challenging task, the current work aimed to assess the performance of various fuzzy models on temperature, solar radiation, and wind speed forecasting. The models studied were taken from the fuzzy systems literature, varying from well-established to the most recent methods. Four cases were considered: a Wang–Mendel (WM)-based fuzzy predictive model, an adaptive network fuzzy inference system (ANFIS), a fuzzy system ensemble, and patch learning (PL). The prediction systems were built from input/output data without any prior information, in a model-free approach. The ability of the models to display high performance on complex real datasets, provided by the National Observatory of Athens, was demonstrated through numerical studies. Patch learning managed to not only display a similar approximation ability to that of strong machine learning models, such as support vector machines and Gaussian processes, but also outperform them on the highly demanding problem of wind speed prediction. More accurately, as far as wind speed prediction is concerned, patch learning produced a 0.9211 root mean squared error for the training data and a value of 0.9841 for the testing data. The support vector machine provided a 0.9306 training root mean squared error and a 0.9891 testing value. The Gaussian process model resulted in a 0.9343 root mean squared error for the training data and a value of 0.9861 for the testing data. Finally, as shown by the numerical experiments, the fuzzy system ensemble exhibited the highest generalisation performance among all the intelligent models.

Enhancing Safety and Efficiency in Firefighting Operations via Deep Learning and Temperature Forecasting Modeling in Autonomous Unit

Firefighters face numerous challenges when entering burning structures to rescue trapped victims, assess the conditions of a residential structure, and extinguish the fire as quickly as possible. These challenges include extreme temperatures, smoke, toxic gases, explosions, and falling objects, which can hinder their efficiency and pose risks to their safety. Accurate information and data about the burning site can help firefighters make informed decisions about their duties and determine when it is safe to enter and evacuate, reducing the likelihood of casualties. This research presents unsupervised deep learning (DL) to classify the danger levels at a burning site and an autoregressive integrated moving average (ARIMA) prediction model to forecast temperature changes using the extrapolation of a random forest regressor. The DL classifier algorithms provide the chief firefighter with an awareness of the danger levels in the burning compartment. The prediction models forecast the rise in temperature from a height ranging from 0.6 m to 2.6 m and the changes in temperature over time at an altitude of 2.6 m. Predicting the temperature at this altitude is critical as the temperature increases faster with height, and elevated temperatures can weaken the building’s structural material. We also investigated a new classification method using an unsupervised DL autoencoder artificial neural network (AE-ANN). The prediction data analytical approach included using the autoregressive integrated moving average (ARIMA) with random forest regression implementation. The proposed AE-ANN model, with an accuracy score of 0.869, did not perform as well compared to previous work, with an accuracy of 0.989, at achieving high accuracy scores for the classification task using the same dataset. However, the random forest regressor and our ARIMA models are analyzed and evaluated in this work, while other research has not utilized this dataset, even though it is open-sourced. However, the ARIMA model demonstrated remarkable predictions of the trends of temperature changes in a burning site. The proposed research aims to classify fire sites into dangerous levels and predict temperature progression using deep learning and predictive modeling techniques. This research’s main contribution is using a random forest regressor and autoregressive integrated moving average models to predict temperature trends in burning sites. This research demonstrates the potential of using deep learning and predictive modeling to enhance firefighter safety and decision-making processes.

Evaluation of High-Resolution Land Cover Geographical Data for the WRF Model Simulations

Increased computing power has made it possible to run simulations of the Weather Research and Forecasting (WRF) numerical model in high spatial resolution. However, running high-resolution simulations requires a higher-detail mapping of landforms, land use, and land cover. Often, higher-resolution data have limited coverage or availability. This paper presents the feasibility of using CORINE Land Cover (CLC) land use and land cover data and alternative high-resolution global coverage land use/land cover (LULC) data from Copernicus Global Land Service Land Cover Map (CGLS-LC100) V2.0 in high-resolution WRF simulations (100 × 100 m). Global LULC data with a resolution of 100 m are particularly relevant for areas not covered by CLC. This paper presents the method developed by the authors for reclassifying land cover data from CGLS-LC100 to MODIS land use classes with defined parameters in the WRF model and describes the procedure for their implementation into the model. The obtained simulation results of the basic meteorological parameters from the WRF simulation using CLC, CGLS-LC100 and default geographical data from MODIS were compared to observations from 13 meteorological stations in the Warsaw area. The research has indicated noticeable changes in the forecasts of temperature, relative humidity wind speed, and direction after using higher-resolution LULC data. The verification results show a significant difference in weather predictions in terms of CLC and CGLS-LC100 LULC data implementation. Due to the fact that better results were obtained for CLC simulations than for CGLS-LC100, it is suggested that CLC data are first used for simulations in numerical weather prediction models and to use CGLS-LC100 data when the area is outside of CLC coverage.

Adaptive thermal model for loading of transformers in low carbon electricity distribution networks

The uptake of low carbon technologies, particularly Electric Vehicles (EVs) and Heat-Pumps (HPs), at the low voltage (LV) distribution network, in the quest of cutting down on greenhouse gas (GHG) emissions in the transportation and residential sectors, has the potential to cause general load increase and may lead to higher and longer peak load demand. This development can, as hinted in previous studies, pose a real challenge of capacity overloading to transformers at the LV distribution network of electricity system. Prolonged periods of transformer overloading could lead to premature transformer failure and shortens transformer's life expectancy. A direct solution to addressing transformer overloading is the upgrading of the transformer capacity. However, the number of LV distribution transformers in electricity system to be upgraded and the resources needed for such operation make the solution less desirable to the Distribution Network Operators (DNOs). Therefore, it is important to develop cost-effective solutions for the optimal utilization of the existing transformer capacity. Adaptive thermal loading of transformers is one of such solutions. This paper focusses on the Adaptive Thermal Loading (ATL) of transformers in LV distribution networks with considerable penetration level of EVs and HPs. The thermal model of a 500-kVA, 11/0.415-kV (no load), 50-Hz, Dyn11, ONAN mineral oil filled, free breathing, ground mounted transformer serving a real and typical urban LV network in the United Kingdom (UK) is developed based on IEC 60,076–7:2005 standard and used as the case study. A method of adaptive thermal loading of the transformer is presented to examine its capacity performance when serving the future load of the LV network following the integration of projected uptake figures of EVs and HPs for the years 2020, 2030, 2040 and 2050 into the network. Given the load and temperature forecasts of a day, the method aims at optimizing, considering the real and present conditions of the operating environment, the overall daily transformer capacity utilization that gives maximum daily return on investment without undermining reliability of supply and normal life expectancy of the transformer. Results show improved performances of the transformer when the adaptive thermal loading method is used.

Using Artificial Neural Networks to Improve CFS Week-3–4 Precipitation and 2-m Air Temperature Forecasts

Abstract Forecast skill from dynamical forecast models decreases quickly with projection time due to various errors. Therefore, postprocessing methods, from simple bias correction methods to more complicated multiple linear regression–based model output statistics, are used to improve raw model forecasts. Usually, these methods show clear forecast improvement over the raw model forecasts, especially for short-range weather forecasts. However, linear approaches have limitations because the relationship between predictands and predictors may be nonlinear. This is even truer for extended range forecasts, such as week-3–4 forecasts. In this study, neural network techniques are used to seek or model the relationships between a set of predictors and predictands, and eventually to improve week-3–4 precipitation and 2-m temperature forecasts made by the NOAA/NCEP Climate Forecast System. Benefitting from advances in machine learning techniques in recent years, more flexible and capable machine learning algorithms and availability of big datasets enable us not only to explore nonlinear features or relationships within a given large dataset, but also to extract more sophisticated pattern relationships and covariabilities hidden within the multidimensional predictors and predictands. Then these more sophisticated relationships and high-level statistical information are used to correct the model week-3–4 precipitation and 2-m temperature forecasts. The results show that to some extent neural network techniques can significantly improve the week-3–4 forecast accuracy and greatly increase the efficiency over the traditional multiple linear regression methods.

Evaluating Atmospheric Surface Layer Flux Parameterization within the Coastal Regime

Abstract Traditional atmospheric surface layer theory assumes homogeneous surface conditions. Regardless, nearly all surface layer parameterization schemes employed within numerical weather prediction models utilize the same techniques within highly heterogeneous coastal regimes as for homogeneous environments. We compare predicted surface weather and fluxes of momentum, heat, and moisture—focusing mainly on momentum—from regional simulations using the Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS) atmospheric model to observations collected from offshore buoys, inland flux towers, and radiosonde profiles during the Coastal Land-Air-Sea Interaction (CLASI) project throughout the summer of 2021 around Monterey Bay, California. Results reveal that modeled cross-coastal surface flux gradients are spuriously discontinuous, leading to systematically overestimated fluxes and weak winds inland of the coastline during onshore flow periods. Additionally, contrary to observations, modeled surface exchange coefficients are insensitive to wind direction on both sides of the coast, which degrades predictive skill downstream from the coastline. Over the central bay, prediction degrades when near-surface wind directions deviate from the prevailing flow direction as the parameterized stress–wind relationship fails during these cases. Predictive skill over the bay is therefore linked to variations in wind direction. Offshore of the geographically complex peninsula, systematic biases are less clear; however, bifurcations in drag coefficients based on wind direction were measured here as well. Last, increasing the horizontal grid spacing from 333 m to 3 km does not significantly affect surface layer prediction. This work highlights the need to reevaluate surface layer parameterization methods for modeling within coastal regions. Significance Statement Understanding surface layer weather is critical for many purposes, such as infrastructure design and weather forecasting. Within the context of numerical modeling and weather prediction, skillful forecasts of surface winds and temperature rely on accurate portrayal of the surface layer. By comparing observations collected during the Coastal Land-Air-Sea Interaction field program to numerical model solutions, we show that prediction of the surface layer fluxes of momentum, heat, and moisture break down near the coastline, which leads to biases in the predicted surface layer weather both inland and over the water. As surface layer parameterization methods across nearly all numerical models are rooted in the same practices, our results call into question the use of traditional methods near the coastline.

The Role of Weather Predictions in Electricity Price Forecasting Beyond the Day-Ahead Horizon

Forecasts of meteorology-driven factors, such as intermittent renewable generation, are commonly included in electricity price forecasting models. We show that meteorological forecasts can be used directly to improve price forecasts multiple days in advance. We introduce an autoregressive multivariate linear model with exogenous variables and LASSO for variable selection and regularization. We used variants of this model to forecast German wholesale prices up to ten days in advance and evaluate the benefit of adding meteorological forecasts, namely wind speed and direction, solar irradiation, cloud cover, and temperature forecasts of selected locations across Europe. The resulting regression coefficients are analyzed with regard to their spatial as well as temporal distribution and are put in context with underlying power market fundamentals. Wind speed in northern Germany emerges as a particularly strong explanatory variable. The benefit of adding meteorological forecasts strongest when autoregressive effects are weak, yet the accuracy of the meteorological forecasts is sufficient for the model to identify patterns. Forecasts produced 2-4 days in advance exhibit an improvement in RMSE by 10-20%. Furthermore, the forecasting horizon is shown to impact the choice of the regularization penalty that tends to increase at longer forecasting horizons.

The Effect of Remote Teleconnection Patterns on Temperature and Precipitation of the Euphrates-Tigris Basin

The Euphrates-Tigris Basin is the most important water source in the Middle East. The present study examined the relationship between the precipitation and temperature characteristics of the basin using remote teleconnection patterns on a monthly time scale. The effects of the North Atlantic, Arctic Oscillations, North Sea Caspian Pattern, and Western Mediterranean indices were examined. The relationship between teleconnection patterns and precipitation/temperature was investigated by adopting Spearman’s Correlation test. All of the remote teleconnection patterns had significant effects on the temperature and precipitation characteristics of the basin. However, the North Sea Caspian Pattern significantly affected the temperatures of the entire basin. Similarly, the Western Mediterranean index had a significant effect on the average temperatures for four months (February, April, November, and December) in almost the entire basin. Also, the Western Mediterranean Index corralates positively with the precipitation of the basin in January, while the correlation is negative in October, and November. Especially, the Western Mediterranean Index and the North Sea Caspian Pattern showed one-month and two-month delayed relationships in monthly total precipitation in some months. At the extremes of the index values, relationships often became strong and distinct. The study results may be useful for seasonal temperature and precipitation forecasts of the Euphrates-Tigris basin.

Integrated adversarial long short-term memory deep networks for reheater tube temperature forecasting of ultra-supercritical turbo-generators

As an important equipment of the ultra-supercritical turbine systems, real-time and accurate control of reheater tube temperature is the key to the normal operation of the system. To improve the accuracy of reheater tube temperature prediction, an integrated adversarial long short-term memory deep network is proposed in this study. Considering the correlation and dynamics between reheater temperature features, different types of single long short-term memory deep networks are batch mixed and stacked in parallel to integrate adversarial training. The training process shares network layer weight parameters to synthesize the decision boundary of each network and complete the dichotomous integration. The proposed networks predict information for the next ten time steps, wirelessly approximate the mapping relationship between multi-source input and output data, and maximize the prediction accuracy and generalization. The proposed networks with single long short-term memory deep networks and 21 algorithms are compared for temperature prediction on the reheater tubes of thermal power plants in Guangdong Province, China, from March 2020 to May 2020. Compared with 90 networks without an integrated adversary, the mean square error obtained by the integrated networks have a 90.85% reduction. Furthermore, compared with the algorithm with the second smallest error among the 22 algorithms, the mean absolute percentage error obtained by the integrated networks is significantly reduced by 99.66%, and the mean absolute error is reduced by 98.36%.

Correction method by introducing cloud cover forecast factor in model temperature forecast

Objective temperature forecast products can achieve better forecast quality by using one-dimensional regression correction directly based on the present model temperature forecast product, and the forecast accuracy can be further improved by adding appropriate auxiliary factors. In this paper, ECMWF forecast products and ground observation data from Fujian are used to revise the surface temperature at 2 m by introducing a cloud cover forecast factor based on the model temperature forecast correction method. Analysis shows that the forecast deviation of daily maximum and minimum temperature after the revision of a single-factor forecast is obviously correlated with cloud cover. A variety of prediction schemes are designed, and the final scheme is determined through comparative testing. The following conclusions are drawn: all schemes based on cloud cover grouping can improve forecast performance, and the total cloud cover scheme is generally better than the low cloud cover scheme. There is a good positive correlation between the forecast deviation of maximum temperature and the mean total cloud cover; that is, the more cloud cover, the bigger the deviation. The minimum temperature is negatively correlated with cloud cover when the cloud cover is less than 40% and positively correlated for the rest. The absolute forecast deviations of the maximum and minimum temperatures are larger when the total cloud cover is less. Whether for Tmax or Tmin forecast, the binary regression scheme after grouping consistently showed the best performance, with the lowest MAE. The final scheme was used to forecast the maximum and minimum temperature in 2021, and most verification indicators showed improvement in most forecast periods. The forecast accuracy for the 36-h daily maximum and minimum temperature is 81.312% and 91.480%, respectively, which is 2.4%–2.6% higher than the single-factor regression scheme. The forecast skill scores (FSS) reach 0.065 and 0.086, indicating that the method can effectively improve forecast quality in a stable manner and can be used for practical forecasting.

Grey-box modeling for hot-spot temperature prediction of oil-immersed transformers in power distribution networks

Power transformers are one of the most costly assets in power grids. Due to increasing electricity demand and levels of distributed generation, they are more and more often loaded above their rated limits. Transformer ratings are traditionally set as static limits, set in a controlled environment with conservative margins. Through dynamic transformer rating, the rating is instead adapted to the actual working conditions of the transformers. This can help distribution system operators (DSOs) to unlock unused capacity and postpone costly grid investments. To this end, real-time information of the transformer operating conditions, and in particular of its hot-spot and oil temperature, is required. This work proposes a grey-box model that can be used for online estimation and forecasting of the transformer temperature. It relies on a limited set of non-intrusive measurements and was developed using experimental data from a DSO in Jutland, Denmark. The thermal model has proven to be able to predict the temperature of the transformers with a high accuracy and low computational time, which is particularly relevant for online applications. With a six-hour prediction horizon the mean average error was 0.4–0.6 °C. By choosing a stochastic data-driven modeling approach we can also provide prediction intervals and account for the uncertainty.

Effect of meteorological data assimilation using 3DVAR on high-resolution simulations of atmospheric CO2 concentrations in East Asia

In this study, the effects of meteorological data assimilation (DA) on high-resolution CO2 concentration simulations in East Asia were investigated using the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) and three-dimensional variational data assimilation (3DVAR). Experiments with and without meteorological DA were performed for June and December 2018. To determine an appropriate DA cycling interval, two experiments were performed at cycling intervals of 6 and 12 h. Compared to the meteorological fields simulated from experiment without DA, those simulated from experiments with DA were more similar to atmospheric observations. The CO2 concentrations simulated from the experiment with a 6 h DA cycling interval were highly similar to the observed CO2 concentrations. Compared to the root mean square error (RMSE) of the experiment without DA, the RMSE of simulated CO2 concentrations in the experiment with a 6 h DA interval decreased by 8.19% in June and 1.43% in December with respect to surface CO2 concentration observations, and decreased by 2.52% in June and 0.47% in December against column-averaged CO2 concentration (XCO2) observations of the Orbiting Carbon Observatory 2 (OCO-2). The improved 2 m temperature forecasts by DA enhanced biogenic CO2 simulations using the Vegetation Photosynthesis and Respiration Model (VPRM) in WRF-Chem. The improved wind speed and direction forecasts by DA led to improved simulations of the horizontal transport of CO2 concentrations and planetary boundary layer height. Overall, enhanced meteorological forecasts by meteorological DA using 3DVAR improved the accuracy of high-resolution CO2 concentration simulations using WRF-Chem in East Asia.

Skill improvement of the yearly updated reforecasts in ECMWF S2S prediction from 2016 to 2022

Hazardous weather events are often accompanied by subseasonal processes, but the forecast skills of subseasonal prediction are still limited. To assess the skill improvement of the constantly updated model version in ECMWF subseasonal-seasonal (S2S) prediction from 2016 to 2022, the performance of yearly updated reforecasts was evaluated against ERA5 reanalysis data using the temporal anomaly correlation coefficient (TCC) as a metric. The newly updated reforecasts exhibit stable superiority at the weather scale of the first two weeks, regardless of whether the 2-m temperature or precipitation forecast is being considered. At the subseasonal time scale starting from the third week, some slight improvements in prediction skills are only found in several tropical regions. Generally, the week-3 TCC values averaged over global land grids still reflect an advancement in prediction skills for updated reforecasts. For the Madden–Julian Oscillation (MJO), reforecasts can reproduce the characteristics of eastward propagation, but there are deviations in the intensity and propagation range of convection anomalies for reforecasts of all seven years. Based on an evaluation of MJO prediction skill using the bivariate anomaly correlation coefficient and bivariate root-mean-square error, some differences are apparent in the MJO prediction skills among the updated reforecasts, but the improvements do not increase monotonically year by year. Despite the inherent limitation of S2S prediction, positive progress has already been achieved via the constantly updated S2S prediction in ECMWF, which reinforces the confidence in further collaboratively improving S2S prediction in the future.摘要在2016年至2022年间, ECMWF次季节预测系统不断升级并逐年完成新的回报试验. 本文考察该预测系统逐年升级带来的预测技巧提升潜力. 从2米气温和降水来看, 在起报之后的前两周内天气尺度上预测技巧表现出逐年稳定提升的趋势; 在从第三周开始的次季节时间尺度上, 预测技巧的提升仅限于热带部分区域. MJO预测技巧并不随着模式升级而逐年单调提升. 尽管目前S2S预测技巧存在局限性, 但目前已有的进展增强了在未来深入合作以提高S2S预测技术的信心.

A Comparison of the Statistical Downscaling and Long-Short-Term-Memory Artificial Neural Network Models for Long-Term Temperature and Precipitations Forecasting

General circulation models (GCMs) run at regional resolution or at a continental scale. Therefore, these results cannot be used directly for local temperatures and precipitation prediction. Downscaling techniques are required to calibrate GCMs. Statistical downscaling models (SDSM) are the most widely used for bias correction of GCMs. However, few studies have compared SDSM with multi-layer perceptron artificial neural networks and in most of these studies, results indicate that SDSM outperform other approaches. This paper investigates an alternative architecture of neural networks, namely the long-short-term memory (LSTM), to forecast two critical climate variables, namely temperature and precipitation, with an application to five climate gauging stations in the Lake Chad Basin. Lake Chad is a data scarce area which has been impacted by severe drought, where water resources have been influenced by climate change and recent agricultural expansion. SDSM was used as the benchmark in this paper for temperature and precipitation downscaling for monthly time–scales weather prediction, using grid resolution GCM output at a 5 degrees latitude × 5 degrees longitude global grid. Three performance indicators were used in this study, namely the root mean square error (RMSE), to measure the sensitivity of the model to outliers, the mean absolute percentage error (MAPE), to estimate the overall performance of the predictions, as well as the Nash Sutcliffe Efficiency (NSE), which is a standard measure used in the field of climate forecasting. Results on the validation set for SDSM and test set for LSTM indicated that LSTM produced better accuracy on average compared to SDSM. For precipitation forecasting, the average RMSE and MAPE for LSTM were 33.21 mm and 24.82% respectively, while the average RMSE and MAPE for SDSM were 53.32 mm and 34.62% respectively. In terms of three year ahead minimum temperature forecasts, LSTM presents an average RMSE of 4.96 degree celsius and an average MAPE of 27.16%, while SDSM presents an average RMSE of 8.58 degree celsius and an average MAPE of 12.83%. For maximum temperatures forecast, LSTM presents an average RMSE of 4.27 degree celsius and an average MAPE of 11.09 percent, while SDSM presents an average RMSE of 9.93 degree celsius and an average RMSE of 12.07%. Given the results, LSTM may be a suitable alternative approach to downscale global climate simulation models’ output, to improve water management and long-term temperature and precipitations forecasting at local level.

Modeling of the land surface temperature as a function of the soil-adjusted vegetation index

Land surface temperature is a physical-environmental variable that is the target of studies in climatology and heat island phenomena resulting from the urbanization model of the 21st century. It is known that each object has a different thermal capacity, which results in higher or lower temperatures. The modeling of temperature as a function of objects on the land surface can allow understanding between these variables. It can corroborate temperature forecasts with the alteration of objects in an area. The objects on the Earth's surface can be computed with geoprocessing techniques that aim to detail Land Use and Occupation. This paper evaluates the Linear, Exponential, and Sinusoidal models to determine which of these models is more expressive for the study of the land surface temperature as a function of land surface objects. For this purpose, images from the Landsat 8 satellite were used to calculate the Earth's surface temperature and determine the scene's objects. To determine the scene's objects, the Soil Adjusted Vegetation Index (SAVI) was used, which is one of the techniques for obtaining Land Use and Occupation. For model analysis, Akaike's Information Criterion (AIC), Bayesian Information Criterion (BIC), and Sum of Square of Residuals (SSR). According to the AIC, BIC, and SSR criteria, the sinusoidal model presented better performance when compared to the other models. However, there are large variations in SSR between classes, especially for the pasture class, which makes the models not highly explanatory.

Variation Characteristics of Temperature and Rainfall and Their Relationship with Geographical Factors in the Qinling Mountains

The Qinling Mountains (QMs) are considered to be the division in geology, geochemistry, and physical geography between northern China and southern China. They have crucial effects on regional climate, especially on rainfall and temperature, and have shown great scientific relevance to climate change research in China. Using the observational daily and monthly rainfall and temperature data derived from meteorological and regional automatic stations—as well as the methods of correlation analysis, climate trend analysis, and Mann–Kendal and t tests—we revealed the spatiotemporal change characteristics of temperature and rainfall and their correlation with elevation, longitude, and latitude. The results show that the annual mean temperature (AMT) underwent a significant increasing trend in the QMs. The maximum AMT increase occurred in spring, and the minimum occurred in summer. Positive anomalies of annual mean rainfall amount (AMRA) occurred in the 1960s, 1980s, and 2010s, and negative anomalies occurred in the 1970s, 1990s, and 2000s. In the QMs, the amount of moderate rainfall (MR) occupied the maximum proportion and accounted for 27.9% of the AMRA, whereas the torrential rainfall (TR) occupied the minimum proportion and accounted for 12.8%. The AMRA amount significantly decreased by 130.1 mm from the 1980s to the 1990s and accounted for 13.5% of the measure in the 1980s. The AMT and AMRA showed consistent change trends with increases in elevation and latitude and showed the opposite trend as the longitude increased. The results offer a further understanding of the meteorological background of the QMs, helping us in further investigating the potential physical mechanisms that influence the spatiotemporal distribution characteristics of temperature and rainfall in the QMs. This study will provide a scientific basis for rainfall and temperature forecasts, with relevance to local ecosystems, agriculture, soil erosion, and the prevention and mitigation of floods in the future.

Comparison between the Operational and Statistical Daily Maximum and Minimum Temperature Forecasts on the Central Coast of Peru

Abstract Multiple linear regression models were developed for 1–3-day lead forecasts of maximum and minimum temperature for two locations in the city of Lima, on the central coast of Peru (12°S), and contrasted with the operational forecasts issued by the National Meteorological and Hydrological Service—SENAMHI and the output of a regional numerical atmospheric model. We developed empirical models, fitted to data from the 2000–13 period, and verified their skill for the 2014–19 period. Since El Niño produces a strong low-frequency signal, the models focus on the high-frequency weather and subseasonal variability (60-day cutoff). The empirical models outperformed the operational forecasts and the numerical model. For instance, the high-frequency annual correlation coefficient and root-mean-square error (RMSE) for the 1-day lead forecasts were 0.37°–0.53° and 0.74°–1.76°C for the empirical model, respectively, but from around −0.05° to 0.24° and 0.88°–4.21°C in the operational case. Only three predictors were considered for the models, including persistence and large-scale atmospheric indices. Contrary to our belief, the model skill was lowest for the austral winter (June–August), when the extratropical influence is largest, suggesting an enhanced role of local effects. Including local specific humidity as a predictor for minimum temperature at the inland location substantially increased the skill and reduced its seasonality. There were cases in which both the operational and empirical forecast had similar strong errors and we suggest mesoscale circulations, such as the low-level cyclonic vortex over the ocean, as the culprit. Incorporating such information could be valuable for improving the forecasts. Significance Statement We wanted to compare the temperature of the operational forecast of the Meteorological and Hydrological Service, an atmospheric model, and persistence with the observed temperatures on the Peruvian central coast. In addition, we generated an empirical forecast model considering both atmospheric and local predictors. We got better results with this empirical model, considering the highest Pearson correlations and the lowest RMSE values. These results will allow us to use this empirical model as the main tool to automate the forecast on the central coast of Peru. Future work should be aimed at testing the skill of this model for forecasting in other cities of Peru.

Evaluating Seasonal Climate Forecasts from Dynamical Models over South America

Abstract Seasonal climate forecasts have socioeconomic value, and the quality of the forecasts is important to various societal applications. Here we evaluate seasonal forecasts of three climate variables, vapor pressure deficit (VPD), temperature, and precipitation, from operational dynamical models over the major cropland areas of South America; analyze their predictability from global and local circulation patterns, such as El Niño–Southern Oscillation (ENSO); and attribute the source of prediction errors. We show that the European Centre for Medium-Range Weather Forecasts (ECMWF) model has the highest quality among the models evaluated. Forecasts of VPD and temperature have better agreement with observations (average Pearson correlation of 0.65 and 0.70, respectively, among all months for 1-month-lead predictions from the ECMWF) than those of precipitation (0.40). Forecasts degrade with increasing lead times, and the degradation is due to the following reasons: 1) the failure of capturing local circulation patterns and capturing the linkages between the patterns and local climate; and 2) the overestimation of ENSO’s influence on regions not affected by ENSO. For regions affected by ENSO, forecasts of the three climate variables as well as their extremes are well predicted up to 6 months ahead, providing valuable lead time for risk preparedness and management. The results provide useful information for further development of dynamical models and for those who use seasonal climate forecasts for planning and management. Significance Statement Seasonal climate forecasts have socioeconomic value, and the quality of the forecasts is important to their applications. This study evaluated the quality of monthly forecasts of three important climate variables that are critical to agricultural management, risk assessment, and natural hazards warning. The findings provide useful information for those who use seasonal climate forecasts for planning and management. This study also analyzed the predictability of the climate variables and the attribution of prediction errors and thus provides insights for understanding models’ varying performance and for future improvement of seasonal climate forecasts from dynamical models.

Predicting indoor air temperature and thermal comfort in occupational settings using weather forecasts, indoor sensors, and artificial neural networks

We present an approach to generate location-specific forecasts of indoor temperature (Ti) and thermal comfort and issue indoor heat warnings for occupational settings. Indoor forecasts are generated using standard outdoor weather forecasting products and an artificial neural network (ANN) trained on-site using local indoor measurements from a low-cost sensor system measuring Ti and indoor physiologically equivalent temperature (PETi). The outcomes are hourly indoor Ti and PETi forecasts. Different ANN-based forecast products using different predictors were concurrently tested at 121 workplaces in agricultural, industrial, storage, and office buildings using data for an entire annual cycle. A forecast was considered skillful when the Ti and PETi forecast was <2 K from actual measurements. The best-performing model used the predictors time of year, week, and day; solar position; and outdoor weather forecast variables to train and run an ANN to predict Ti or PETi. It had an annual average mean absolute forecast error of 0.87 K for Ti and 0.99 K for PETi over the next 24 h, with Pearson correlation coefficients of 0.98 and 0.97, respectively. Overall, 91% of Ti forecasts and 88% of PETi forecasts were skillful. Indoor forecasts showed larger errors in the summer than in the winter. We conclude that combining indoor data with weather forecasts using ANNs could be implemented widely to provide location-specific indoor weather forecasts to improve and localize heat and health warning systems.