It is often claimed that memory for weather events is short, resulting in limited retention of extreme events, which we tend to exaggerate. There is a prevailing belief among people of all ages that winters have changed significantly from their historical patterns. But is this belief correct? Is it true that winters are becoming shorter in Spain?
In this article we are going to analyze If winters are getting shorter in Spain.
Cold days in Spain
For many years, a lower than anticipated frequency of cold days has been documented, attributed to basic natural variability, along with generally high minimum temperatures, in line with the forecasts outlined by several climate change scenarios. One of them predicts a decrease in the duration of winters and an increase in the duration of summers. A study by César Rodríguez examined temperature records from several observatories across the country and found that summers are lengthening by 4 to 15 days per decade, depending on the region. Based on this research, a similar analysis was carried out throughout the country. This current study aims to carry out the same assessment specifically for winter.
What methodology was used to perform the calculations?
To determine potential winter disturbances, the initial task is to establish a definition of winter and its duration. The analysis has been carried out using the Copernicus Climate Change Service (C5S) ERA3 reanalysis database, which provides hourly data from 1940 to the present, with a spatial resolution of 0,25⁰ for both latitude and longitude. The subsequent calculations have been carried out for each specific point:
The average daily temperature has been determined for the period between December 1st and February 28th for the Peninsula and the Balearic Islands, or between December 15th and March 15th for the Canary Islands, covering the years 1991 to 2020. To avoid including non-winter periods that had minimum temperatures characteristic of that season but that deviated significantly from typical winter conditions (including low maximum temperatures), the average temperature was used instead of the minimum temperature.
Thus, Periods, often indicative of spring, which could be misleading and unnecessarily lengthen the winter season, have been effectively excluded. A threshold based on the 70th percentile of the distribution of daily mean temperatures was then established, allowing outlier days with significantly high positive anomalies to be removed. The threshold values at each location are illustrated in the image below.
For each year from 1940 to 2022, the nine-day moving average temperature was calculated and the beginning and end days of the year when the designated temperature threshold was reached or fell below were identified. To avoid the occurrence of false positives, A prolonged winter period was chosen rather than focusing on an individual spring or autumn day characterized by an unusually low temperature.The temperature of a specific day was determined by averaging the values of the four days before and the four days after, resulting in a total of nine days.
Finally, the dates of the 82 winters were examined for increasing or decreasing trends by applying the Mann-Kendall test. Although the methodology and thresholds may seem subjective, the results demonstrated consistent and comparable behavior when certain parameters were modified (including percentiles, dates, climatological period, etc.), suggesting that winter length is indeed decreasing. However, it is critical to recognize that the findings should be considered an approximation and not definitive or rigid values.
Winter decline
Studies reveal a notable reduction in the winter season in almost all territories since the 1940s. This decline is less pronounced in the southwestern region of the peninsula, while the central and eastern areas show a more significant decrease, with many localities experiencing a reduction of more than one month. On average, The length of winters today is one month shorter than in the mid-20th century..
There are regions where the trend cannot be confirmed with more than 95% certainty (p-value greater than 0,05). In percentage terms, Winter has shortened by more than 30% in more than half of the country. It is important to note that this important trend is also observed in the maritime regions of the Canary Islands, where the reduction is even more pronounced and clear than in the Peninsula. Part of the explanation for this phenomenon may lie in the lower temperature amplitudes experienced in the archipelago; as a result, even minimal variations in temperature can cause substantial changes in the seasonal calendar.
It is essential to recognize that a downward trend does not suggest that all winters are shorter than their predecessor. The past ten years have seen winters of varying length, but if one looks at a broader time frame, An increasing prevalence of shorter winters is observed along with a reduction in their total duration.
Winter and Autumn
The length of winters is being reduced as a result of the "theft" of both spring and autumn. This gradual lengthening It is not uniform, generally showing a more significant increase in spring than in autumn. The lengthening of spring, known as "invernavera", is especially evident in the southern, central and eastern regions of the peninsula, where in many areas there is an increase of more than three weeks. Furthermore, this trend is observed in a large part of the territory. A similar pattern is observed in the Canary Islands. Thus, it can be concluded that the shortening of spring due to the lengthening of summer is compensated by its lengthening at the expense of winter.
As for the phenomenon of winter reduction by autumn, called "inverotoño", the magnitude is minimal. In the southern regions and some eastern areas, the reduction is remarkably low, with a level of certainty that does not exceed 95%. In other regions, however, the trend is longer than ten days and the certainty is quite clear. In particular, in the Canary Islands, autumn is longer in the western areas, lasting more than three weeks. The results of this analysis invite speculation about the potential duration of winter in 82 years; however, such projections must be approached with great caution.
The gradual decline of winter is characterized by non-linear and non-monotonous patterns, showing periods of pronounced trends, stagnation and even slight increases. In general, a downward trend is observed in the Peninsula since the 1980s and 1990s, which corresponds to the second half of the period studied. On the other hand, The Canary Islands did not experience this trend until the mid-1990s. It is therefore plausible that winter decline is accelerating.
The evidence is consistent with the view that winters are becoming shorter, a trend that is in line with projections related to climate change. This decrease in winter length is somewhat offset by a lengthening of spring, while the effects observed in autumn are less significant.