Meteorologists, sunspotters and journalists: the demise of long-range weather forecasting in the USSR, 1970s.

People generally use the terms weather and climate for the same phenomenon, without realizing the differences between the two. The weather at a place is defined by measuring certain atmospheric parameters such as temperature, pressure, humidity, wind strength and direction, rainfall, snowfall, cloud, sunshine, etc, at a particular time. Depending on the values of these parameters, the weather of a place may be hot or cold, dry or humid, clear or cloudy, rainy, and so on. There are various ways of defining the state of the weather at a place, at a particular time. Weather forecasts are normally for a specific place, and are valid for a specific period of time. For example, the weather forecast for up to 2–3 days is known as a short-range forecast; that for a period of 3–10 days is known as a medium-range weather forecast, and that for a month or season is known as a long-range or extendedrange weather forecast. These forecasts provide the state of the atmosphere during the stated number of days, in terms of the temperature, surface pressure, humidity, rainfall, snowfall, clouds, sunshine, fog, frost, thunder, gale, and so on. These forecasts also state the possibility of severe weather conditions, such as cyclones, tornadoes, thunderstorms, floods, droughts, avalanches, and so on. Every country has its own operational weather agency, which regularly monitors the state of the atmosphere at different weather observatories.

Cyclicality is a phenomenon commonly observed in nature, often in relation to natural events, including weather. Can cyclicality be used, however, to accurately and reliably predict long-range weather? For the past 30-plus years what appears to be a regularly cycling pattern has been investigated, researched, and tested in the development of a forecast system designed to make weather predictions using knowledge of this cycling pattern. Going back further, this may have been discovered as early as the 1940s. Long-range weather forecasting using this methodology is currently being utilized with accurate predictions of weather that is experienced at the surface from the next day to a likely limit of 300 days into the future. This method introduced in this study has demonstrable accuracy and robustness from December to September within a given forecast year. If this hypothesis of cyclicality plays an important role in weather forecasting, this seminal methodology represents a paradigm shift from current weather forecasting methods. Specific examples of cyclicality in the 500-hPa height fields from the 2016-2017 season will be showcased. For example, it will be shown how the 500-hPa height fields and surface weather can be accurately predicted months in advance based on how the weather pattern set up and cycled in the early fall. Specifically, this Cycling Pattern Hypothesis will be applied to the potential cyclicality of extreme precipitation events in the Lake Tahoe, NV (USA) area during the drought ending 2016-2017 season over the western United States. This new hypothesis may provide answers and solutions to forecasting droughts, floods, and more.

Abstract. Long-range weather forecasts provide predictions of atmospheric, ocean and land surface conditions that can potentially be used in land surface and hydrological models to predict the water and energy status of the land surface or in crop growth models to predict yield for water resources or agricultural planning. However, the coarse spatial and temporal resolutions of available forecast products have hindered their widespread use in such modelling applications, which usually require high-resolution input data. In this study, we applied sub-seasonal (up to 4 months) and seasonal (7 months) weather forecasts from the latest European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system (SEAS5) in a land surface modelling approach using the Community Land Model version 5.0 (CLM5). Simulations were conducted for 2017–2020 forced with sub-seasonal and seasonal weather forecasts over two different domains with contrasting climate and cropping conditions: the German state of North Rhine-Westphalia (DE-NRW) and the Australian state of Victoria (AUS-VIC). We found that, after pre-processing of the forecast products (i.e. temporal downscaling of precipitation and incoming short-wave radiation), the simulations forced with seasonal and sub-seasonal forecasts were able to provide a model output that was very close to the reference simulation results forced by reanalysis data (the mean annual crop yield showed maximum differences of 0.28 and 0.36 t ha−1 for AUS-VIC and DE-NRW respectively). Differences between seasonal and sub-seasonal experiments were insignificant. The forecast experiments were able to satisfactorily capture recorded inter-annual variations of crop yield. In addition, they also reproduced the generally higher inter-annual differences in crop yield across the AUS-VIC domain (approximately 50 % inter-annual differences in recorded yields and up to 17 % inter-annual differences in simulated yields) compared to the DE-NRW domain (approximately 15 % inter-annual differences in recorded yields and up to 5 % in simulated yields). The high- and low-yield seasons (2020 and 2018) among the 4 simulated years were clearly reproduced in the forecast simulation results. Furthermore, sub-seasonal and seasonal simulations reflected the early harvest in the drought year of 2018 in the DE-NRW domain. However, simulated inter-annual yield variability was lower in all simulations compared to the official statistics. While general soil moisture trends, such as the European drought in 2018, were captured by the seasonal experiments, we found systematic overestimations and underestimations in both the forecast and reference simulations compared to the Soil Moisture Active Passive Level-3 soil moisture product (SMAP L3) and the Soil Moisture Climate Change Initiative Combined dataset from the European Space Agency (ESA CCI). These observed biases of soil moisture and the low inter-annual differences in simulated crop yield indicate the need to improve the representation of these variables in CLM5 to increase the model sensitivity to drought stress and other crop stressors.

<strong class="journal-contentHeaderColor">Abstract.</strong> Long-range weather forecasts provide predictions of atmospheric, ocean and land surface conditions that can potentially be used in land surface and hydrological models to predict the water and energy status of the land surface or in crop growth models to predict yield for water resources or agricultural planning. However, the coarse spatial and temporal resolutions of available forecast products have hindered their widespread use in such modelling applications that usually require high resolution input data. In this study, we applied sub-seasonal (up to 4 months) and seasonal (7 months) weather forecasts from the latest European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system (SEAS5) in a land surface modelling approach using the Community Land Model version 5.0 (CLM5). Simulations were conducted for 2017&ndash;2020 forced with sub-seasonal and seasonal weather forecasts over two different domains with contrasting climate and cropping conditions: the German state of North Rhine-Westphalia and the Australian state of Victoria. We found that, after pre-processing of the forecast products (temporal downscaling of precipitation and incoming shortwave radiation), the simulations forced with seasonal and sub-seasonal forecasts were able to generate a model system response very close to reference simulation results forced by reanalysis data. Differences between seasonal and sub-seasonal experiments were insignificant. The forecast experiments were able to satisfactorily capture recorded inter-annual variations of crop yield. In addition, they also reproduced the generally higher inter-annual variability in crop yield across the Australian domain (approximately 50 % inter-annual variability in recorded yields and up to 17 % in simulated yields) compared to the German domain (approximately 15 % inter-annual variability in recorded yields and up to 5 % in simulated yields). The high and low yield seasons (2020 and 2018) among the four simulated years were clearly reproduced in forecast simulation results. Furthermore, sub-seasonal and seasonal simulations reflected the early harvest in the drought year of 2018 in the German domain. However, the simulated inter-annual yield variability was lower in all simulations compared to the official statistics. While general soil moisture trends, such as the European drought in 2018, were captured by the seasonal experiments, we found systematic over- and underestimations in both the forecast and the reference simulations compared to the Soil Moisture Active Passive Level-3 soil moisture product (SMAP L3) and the Soil Moisture Climate Change Initiative Combined dataset from the European Space Agency's (ESA CCI). These observed biases of soil moisture as well as the low inter-annual variability of simulated crop yield indicate the need to improve the representation of these variables in CLM5 to increase the model sensitivity to drought stress and other crop stressors.

Modern life depends on precise weather forecasting because global warming intensifies which affects how people live while handling energy needs and maintaining agriculture and protecting the environment. This study proposes a temperature data prediction system which integrates convolutional neural networks (CNNs) with long short-term memory networks (LSTMs). The CNN-LSTM hybrid model connects two network types to process time sequences and detect spatial information alike. The hybrid CNN-LSTM combines temporal and spatial processing to generate weather forecasts which are dependable and precise for meteorological data analysis. Researchers confirm that adding CNN-LSTM technology increases prediction accuracy especially for intricate tasks such as long-range weather forecasting. The combination of CNN and LSTM models brings strong performance in weather forecasting due to its success handling large and diverse meteorological data types. Time-dependent data management through LSTMs produces highly accurate and stable predictions while spatial feature extraction relies on CNNs. During processing of complex meteorological information, the model demonstrates excellent performance by handling problems related to data dimensions and missing values. MAE functions as the chosen loss function in this model. The testing results prove the potential of this climatology prediction model through its ability to produce curves that match test data measurement results. This research establishes essential foundations for future weather prediction systems within global climate change scenarios and provides valuable findings that benefit agriculture as well as energy management and urban development practices.

<strong class="journal-contentHeaderColor">Abstract.</strong> Long-range weather forecasts provide predictions of atmospheric, ocean and land surface conditions that can potentially be used in land surface and hydrological models to predict the water and energy status of the land surface or in crop growth models to predict yield for water resources or agricultural planning. However, the coarse spatial and temporal resolutions of available forecast products have hindered their widespread use in such modelling applications that usually require high resolution input data. In this study, we applied sub-seasonal (up to 4 months) and seasonal (7 months) weather forecasts from the latest European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system (SEAS5) in a land surface modelling approach using the Community Land Model version 5.0 (CLM5). Simulations were conducted for 2017&ndash;2020 forced with sub-seasonal and seasonal weather forecasts over two different domains with contrasting climate and cropping conditions: the German state of North Rhine-Westphalia and the Australian state of Victoria. We found that, after pre-processing of the forecast products (temporal downscaling of precipitation and incoming shortwave radiation), the simulations forced with seasonal and sub-seasonal forecasts were able to generate a model system response very close to reference simulation results forced by reanalysis data. Differences between seasonal and sub-seasonal experiments were insignificant. The forecast experiments were able to satisfactorily capture recorded inter-annual variations of crop yield. In addition, they also reproduced the generally higher inter-annual variability in crop yield across the Australian domain (approximately 50 % inter-annual variability in recorded yields and up to 17 % in simulated yields) compared to the German domain (approximately 15 % inter-annual variability in recorded yields and up to 5 % in simulated yields). The high and low yield seasons (2020 and 2018) among the four simulated years were clearly reproduced in forecast simulation results. Furthermore, sub-seasonal and seasonal simulations reflected the early harvest in the drought year of 2018 in the German domain. However, the simulated inter-annual yield variability was lower in all simulations compared to the official statistics. While general soil moisture trends, such as the European drought in 2018, were captured by the seasonal experiments, we found systematic over- and underestimations in both the forecast and the reference simulations compared to the Soil Moisture Active Passive Level-3 soil moisture product (SMAP L3) and the Soil Moisture Climate Change Initiative Combined dataset from the European Space Agency's (ESA CCI). These observed biases of soil moisture as well as the low inter-annual variability of simulated crop yield indicate the need to improve the representation of these variables in CLM5 to increase the model sensitivity to drought stress and other crop stressors.

<strong class="journal-contentHeaderColor">Abstract.</strong> Long-range weather forecasts provide predictions of atmospheric, ocean and land surface conditions that can potentially be used in land surface and hydrological models to predict the water and energy status of the land surface or in crop growth models to predict yield for water resources or agricultural planning. However, the coarse spatial and temporal resolutions of available forecast products have hindered their widespread use in such modelling applications that usually require high resolution input data. In this study, we applied sub-seasonal (up to 4 months) and seasonal (7 months) weather forecasts from the latest European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system (SEAS5) in a land surface modelling approach using the Community Land Model version 5.0 (CLM5). Simulations were conducted for 2017&ndash;2020 forced with sub-seasonal and seasonal weather forecasts over two different domains with contrasting climate and cropping conditions: the German state of North Rhine-Westphalia and the Australian state of Victoria. We found that, after pre-processing of the forecast products (temporal downscaling of precipitation and incoming shortwave radiation), the simulations forced with seasonal and sub-seasonal forecasts were able to generate a model system response very close to reference simulation results forced by reanalysis data. Differences between seasonal and sub-seasonal experiments were insignificant. The forecast experiments were able to satisfactorily capture recorded inter-annual variations of crop yield. In addition, they also reproduced the generally higher inter-annual variability in crop yield across the Australian domain (approximately 50 % inter-annual variability in recorded yields and up to 17 % in simulated yields) compared to the German domain (approximately 15 % inter-annual variability in recorded yields and up to 5 % in simulated yields). The high and low yield seasons (2020 and 2018) among the four simulated years were clearly reproduced in forecast simulation results. Furthermore, sub-seasonal and seasonal simulations reflected the early harvest in the drought year of 2018 in the German domain. However, the simulated inter-annual yield variability was lower in all simulations compared to the official statistics. While general soil moisture trends, such as the European drought in 2018, were captured by the seasonal experiments, we found systematic over- and underestimations in both the forecast and the reference simulations compared to the Soil Moisture Active Passive Level-3 soil moisture product (SMAP L3) and the Soil Moisture Climate Change Initiative Combined dataset from the European Space Agency's (ESA CCI). These observed biases of soil moisture as well as the low inter-annual variability of simulated crop yield indicate the need to improve the representation of these variables in CLM5 to increase the model sensitivity to drought stress and other crop stressors.

<strong class="journal-contentHeaderColor">Abstract.</strong> Long-range weather forecasts provide predictions of atmospheric, ocean and land surface conditions that can potentially be used in land surface and hydrological models to predict the water and energy status of the land surface or in crop growth models to predict yield for water resources or agricultural planning. However, the coarse spatial and temporal resolutions of available forecast products have hindered their widespread use in such modelling applications that usually require high resolution input data. In this study, we applied sub-seasonal (up to 4 months) and seasonal (7 months) weather forecasts from the latest European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system (SEAS5) in a land surface modelling approach using the Community Land Model version 5.0 (CLM5). Simulations were conducted for 2017&ndash;2020 forced with sub-seasonal and seasonal weather forecasts over two different domains with contrasting climate and cropping conditions: the German state of North Rhine-Westphalia and the Australian state of Victoria. We found that, after pre-processing of the forecast products (temporal downscaling of precipitation and incoming shortwave radiation), the simulations forced with seasonal and sub-seasonal forecasts were able to generate a model system response very close to reference simulation results forced by reanalysis data. Differences between seasonal and sub-seasonal experiments were insignificant. The forecast experiments were able to satisfactorily capture recorded inter-annual variations of crop yield. In addition, they also reproduced the generally higher inter-annual variability in crop yield across the Australian domain (approximately 50 % inter-annual variability in recorded yields and up to 17 % in simulated yields) compared to the German domain (approximately 15 % inter-annual variability in recorded yields and up to 5 % in simulated yields). The high and low yield seasons (2020 and 2018) among the four simulated years were clearly reproduced in forecast simulation results. Furthermore, sub-seasonal and seasonal simulations reflected the early harvest in the drought year of 2018 in the German domain. However, the simulated inter-annual yield variability was lower in all simulations compared to the official statistics. While general soil moisture trends, such as the European drought in 2018, were captured by the seasonal experiments, we found systematic over- and underestimations in both the forecast and the reference simulations compared to the Soil Moisture Active Passive Level-3 soil moisture product (SMAP L3) and the Soil Moisture Climate Change Initiative Combined dataset from the European Space Agency's (ESA CCI). These observed biases of soil moisture as well as the low inter-annual variability of simulated crop yield indicate the need to improve the representation of these variables in CLM5 to increase the model sensitivity to drought stress and other crop stressors.

The focus of this paper is on the problem of airline schedule recovery when airspace capacity is limited by convective weather. The problem is represented in a simulation, the Boeing National Flow Model (NFM), and an analysis is conducted to measure the benefits of increased automation and improved weather forecasting. In particular, an analysis is undertaken based upon all scheduled traffic and the actual convective weather experienced on specific good and bad weather days for the U.S. National Airspace System (NAS) in the year 2007. The analysis quantifies the benefits associated with a collaborative flow management concept involving distributed airline schedule recovery. The concept permits airline users to optimize or re-plan their own schedules given an assigned and equitable share of forecast system capacities. The methodology can be used to evaluate alternative convective weather forecasting products, including both polygonal forecasts (such as the CCFP), along with gridded probabilistic forecasts. The methodology can also utilize the actual weather, typically available in snapshots every 5-6 minutes, as a gridded 0/1 “perfect” forecast. The analysis was conducted and is reported on in two major phases. In both phases of the study we allowed the use of strategies for ground delay and flight cancellation but did not allo w for re-routing. The first phase of the analysis was an initial attempt to study the relative value of alternative forecasts and levels of planner automation and utilized multiple forecasting products in their original form. In this phase, we discovered a systematic difference (i.e., bias) in overall coverage between the forecast and actual weather. The forecasts contained much higher levels of weather coverage, leading to re-planning responses that were considerably over-reactive. These findings led to a forecast calibration effort at NOAA that served to adjust both the forecast and actual weather coverage to greatly reduce or eliminate the bias. In the second phase of the study we undertook to compare multiple calibrated forecasting products as well as alternative levels of planner automation.

Long-range weather forecasts represent an important tool for planning across various sectors, from agriculture to energy. Their use on a global scale continues to grow, supported by the improving quality of numerical models. However, under the conditions of Central Europe—particularly in the Czech Republic—their reliability and applicability face a number of challenges. This contribution focuses on the potential of statistical postprocessing of long-range air temperature forecasts in the Czech Republic.Dynamical forecasts use full three-dimensional climate models to simulate potential changes in the atmosphere and oceans over the coming months based on current conditions. Ensembles of simulations provide probabilistic weather scenarios that indicate the likelihood of a given period being wetter, drier, warmer, or colder compared to the seasonal average. The added value of various postprocessing approaches for seasonal forecasts remains a topic of ongoing debate.This work focuses on statistical postprocessing based on empirical relationships derived between a locally observed predictand of interest (in this case, air temperature) and one or more suitable model predictors from global seasonal forecasting systems. The study analyses the seasonal forecast systems available in the Copernicus Climate Change Service (C3S) archive, which provide near-surface air temperature data at 1° × 1° spatial resolution. It examines the statistical postprocessing of air temperature forecasts for the Czech Republic using four weather forecast systems: the European Centre for Medium-Range Weather Forecasts (ECMWF), Météo-France (MF), Deutscher Wetterdienst (DWD), and Centro Euro-Mediterraneo sui Cambiamenti Climatici (CMCC).The analysis covers the period 1993–2016, which represents the longest hindcast period common to all systems, and the domain of the Czech Republic in Central Europe (49–51°N, 12–19°E). For statistical postprocessing using a neural network method implemented in STATISTICA software, air temperature and sea level pressure data from global forecast models were used as predictors. The reference data used in this study are gridded station-based observational air temperature datasets. The forecast performance is evaluated across three temperature categories: above normal, normal, and below normal.

IN a series of questions asked by Mr. S. F. Markham in the House of Commons on June 22, relating to the stations and research staff of the Meteorological Office, one referred to weather forecasts for a fortnight or longer, now being published in Great Britain and in Germany, and suggested that the Office might supplement its present forecasts by such long-range predictions. In his reply, the Secretary of State for Air said: “I am aware of the long-range weather forecasts being attempted by various methods in many different countries. These efforts are being carefully studied by the Meteorological Office, but so far none of the methods has attained the accuracy which would justify the issue of such forecasts in this country”. This reply may not satisfy the public, which fails entirely to distinguish between weather forecasts based upon established scientific principles and observations, from prophecies of an astrological nature or any system which has not been submitted to a scientific society for disinterested consideration. Whatever is known about long-range weather forecasting is understood by our Meteorological Office, and if any practical end could be served by applying such knowledge, advantage would certainly be taken of the opportunity. In science it is not enough for an observer to satisfy himself that his investigations prove a principle, but the evidence has to convince other scientific workers before the principle is accepted. Until this has been done, any long-range weather forecasts published in the daily Press, whatever accuracy is claimed for them, are altogether unworthy of being placed in the same category as the daily forecasts at present issued.

The feasibility of obtaining spaceborne observations of the oceans has been demonstrated through the flight of NASA's Seasat and Nimbus-7 satellites. As a consequence, ocean-related spacecraft are under various stages of development in Canada, Europe, Japan, and the United States. These spacecraft, a number of which are planned for launch around 1990, will provide a capability for obtaining improved oceanic observations globally. This will enable significant advances in marine forecasting and climatology, which in turn can contribute to the efficiency of marine operations and the design of offshore structures. INTRODUCTION The first oceanic data collected by satellites were the visual and photographic observations made by astronauts in the Mercury Program two decades ago. Since that time, oceanic sensors have evolved from the infrared radiometers on the early meteorological satellites, to the microwave instrument on Skylab in 1973, and finally to the suite of microwave sensors on Seasat and the color scanner on Nimbus-7, both launched in 1978. Given the experience gained in working with the data from these experimental satellites, there now exists a good understanding of sensor performance and the algorithms required to convert raw spacecraft data into geophysical useful information. For example, a demonstrated capability now exists for determining surface winds, waves, temperature, and currents; ocean color; and sea ice cover and motion. With the exception of ocean color and high-spatial-resolution temperature, these can be determined under all-weather conditions. Wilson (1981) provides a brief overview of this capability; Brown and Cheney (1983) give a comprehensive introduction to the literature; and Maul (1985) and Stewart (1985) are the first two textbooks covering this field. FUTURE DIRECTIONS As a consequence of the capability demonstrated by these experimental satellites, there are a number of ocean-related missions both for research and operational purposes under various stages of development and planned for launch around 1990. Simmons (1985) gives an outline description of these missions and their current status. For research, the thrust of NASA's Ocean Topography Experiment (TOPEX) and the NASA Scatter meter (NSCAT) for flight aboard the Navy Remote Ocean Sensing System (NROSS) is to better understand how atmospheric winds drive the circulation of the oceans, and how the oceans in turn have a feedback on the atmosphere via phenomena like the EI Nino. With NASA and the National Science Foundation taking the lead, such an understanding will enable improvements in long-range weather and climate forecasting. Baker (1984) provides a detailed overview of an ocean research strategy for the decade involving space borne observations. On the other hand, the thrust of the overall NROSS mission is to provide global all-weather oceanic observations for operational use by the U. S. Navy and the National Oceanic and Atmospheric Administration (NOAA). The accuracy of marine forecasts is now limited by the availability of good surface observations.

Sea surface temperature (SST) is considered a strong indicator of climate change, being an essential parameter for long-range weather and climate forecasting. Another important indicator of climate change is sea level (SL), which has a longer history of systematic instrumental observations. This paper aims to examine the relationships between low-latitude variations in ocean characteristics (SST and SL) and surface air temperature (SAT) anomalies in the Arctic and mid-latitudes, and discuss the possibility of using SST and SL as predictors to forecast seasonal SAT anomalies. Archives of meteorological observations, atmospheric and oceanic reanalyses, and long-term series of tide gauge data on SL were used in this study. An analysis of relationships between seasonal SAT in different mid-to-high latitude regions and SST made it possible to identify areas in the ocean that have the greatest influence on SAT patterns. The most commonly identified area is located in the tropical North Atlantic. Another area was found in the Indo-Pacific warm pool. The predictive potential of the relationships identified between ocean characteristics (SST and SL) and SAT will be used to build deep learning models aimed at predicting climate variability in mid-to-high latitudes.

This study provides a preliminary assessment of the potential for determining probabilities of future natural-gas-supply interruptions by combining long-range weather forecasts and natural-gas supply/demand projections. An illustrative example which measures the probability of occurrence of heating-season natural-gas curtailments for industrial users in the southeastern US is analyzed. Based on the information on existing long-range weather forecasting techniques and natural gas supply/demand projections enumerated above, especially the high uncertainties involved in weather forecasting and the unavailability of adequate, reliable natural-gas projections that take account of seasonal weather variations and uncertainties in the nation's energy-economic system, it must be concluded that there is little possibility, at the present time, of combining the two to yield useful, believable probabilities of heating-season gas curtailments in a form useful for corporate and government decision makers and planners. Possible remedial actions are suggested that might render such data more useful for the desired purpose in the future. The task may simply require the adequate incorporation of uncertainty and seasonal weather trends into modeling systems and the courage to report projected data, so that realistic natural gas supply/demand scenarios and the probabilities of their occurrence will be available to decision makers during a time when such information is greatly needed.

A brief account is given of Baur's long-range weather forecast prepared in the autumn of 1941 for the 1941–42 winter in Eastern Europe. Baur's forecast called for a ‘normal’ or mild winter but the winter turned out to be one of the most severe winters on record. The cold, the icy winds and blizzards gravely hit the German armies and coincided with the first major Soviet counteroffensive of the war. A Soviet weather forecast for January 1942, also called for a mild month. A review of the climatological studies prepared for the war indicates that the occurrence of mud periods of considerable intensity in autumn was not considered. The autumn 1941 mud period immobilized most of the German armies for a month and caused the attempted final German assault on Moscow to take place in an early and severe winter. Hitler would not tolerate the mention of winter and still less the mention of the retreat of Napoleon's Grande Armee from Russia. The support given by Soviet meteorologists and hydrologists to the Red Army...