Introduction

Climate change is one of the most topical problems worldwide. According to the Intergovernmental Panel on Climate Change (^{IPCC, 2007}), in the past 100 years the global temperature has increased by 0.74 °C. ^{Yu et al. (2015)} mention that in the last 30 years the warm periods were higher than the average, as a result it is expected that the frequency and intensity of extreme temperatures increase. The impact of extreme weather occurs largely in natural disasters such as flooding, heat waves and fire.

This impact happens regularly when extreme weather occurs with some regularity and intensity, causing ecological disasters and negatively affecting society (^{Degefie et al., 2014}). One of the main causes of the increase in average temperature on the planet is the emission of greenhouse gases from human activities (^{IPCC, 2007}). The ocean warming, increased continental temperatures, extreme temperature and behavior patterns of the winds are influenced by this phenomenon and can result in unprecedented extreme climate (IPCC, 2012).

Furthermore, the biogeochemical processes as the case of carbon and nitrogen cycle can be controlled by heat waves, which can change the function of these processes affecting the ecosystem (^{Degefie et al., 2014}). ^{Arnone et al. (2008)} observed that the increase in the frequency of warm years decreased CO2 uptake from terrestrial ecosystem. Studies in different parts of the world using simulation models related to weather events, confirm the trend of more frequent and severe extreme events in the future (^{Schar et al., 2004}; ^{Meehl et al., 2005}; ^{Tebaldi et al., 2006}; ^{Koutroulis et al., 2013}). Other recent studies report a decrease in the frequency of extreme cold and increase of temperature both day and night (^{Acquaotta et al., 2010}; ^{Wang et al., 2013}).

It is also reported that in the last 50 years has decreased the number of cold nights and the number of warm nights is increasing in 70% of the planet (^{Alexander et al., 2006}).

In the United States, ^{Griffiths and Bradley (2007)} studied the indicators of changes in extreme weather conditions in the Northeast region, noticing an increasing trend of heat waves. Furthermore, the ^{IPCC (2007)} mentioned Mexico's vulnerability to climate change, noting that in northwest Mexico and southwest United States, the problem of water scarcity may be exacerbated due to increased temperature and reduced precipitation (^{Cavazos and Arriaga-Ramirez, 2012}). According to ^{Magaña et al. (2012)} regions with greatest risk of increased average annual temperature are expected to occur in northeastern and northwestern Mexico (^{Magaña et al., 2012}).

However, several studies on extreme temperatures in Mexico, mentioned that after 1970, there was a significant increase in the maximum temperature compared to minimum temperatures, especially in Northwest and Central regions of Mexico, which showed contrasting trends regarding the rest of the country (^{Englehart and Douglas, 2005}; ^{Pavia et al., 2008}; ^{Gutierrez-Ruacho et al., 2010}).

Studies made by ^{Peralta-Hernández et al. (2009)} in southeastern Mexico, showed significant trend in the increase of the maximum temperatures in the region, as well as in the frequency of warm days. Despite the importance of this issue, there is no information on trends from climate change in indices in the state of Chiapas. Therefore, the objective of this study was to determine the trends of climate indices, based on maximum temperature, minimum temperature and precipitation recorded in the meteorological stations in the state of Chiapas, for the period 1960-2009.

Materials and methods

Description of the study area

The state of Chiapas is located south of Mexico, bordered to the north by Tabasco, to the south by the Pacific Ocean, on the east by the Republic of Guatemala and to the west by the states of Oaxaca and Veracruz (Figure 1). It is located between 14° 39' and 17° 57' north latitude and 90° 22' and 94° 08' west longitude, its geography is variable, has high mountains, plateaus, mountains, hills, valleys and coasts. The altitude in the state of Chiapas ranges from 0 m on the border with the sea, to a little more than 4 000 m in the Tacana Volcano. The climatic types of largest land area are warm tropical and warm humid tropical, but are also represented other climatic variations such as humid subtropical temperate and semi-warm humid subtropical, among others (^{Medina et al., 1998}).

Figure 1 Location of the study area. 

For this study the information from 16 weather stations spread across the state of Chiapas, with records for the years 1960 to 2009 were used. For data analysis considered the following criteria: 30 years or more of information; 95% or more data; removal of years with 3 months or more without data; removal of weather stations with several consecutive years without data. Figure 2 shows the location of weather stations from which information was analyzed.

Figure 2 Location of the weather stations studied. 

Quality and homogeneity of series

Data was subjected to quality control, using the RClimDex v. 1.0 module, implemented in R v. 3.0.1. Using this module, correction and deletion of data was made after comparing them with the values of neighboring stations on the same dates (Klein-Tank et al., 2009; ^{Zhang and Feng, 2004}). Homogeneity analysis and ruptures detection in the series of analyzed indices was made according to ^{Wang (2008)}; F test maximum penalty using the RHTest v. 3.0 module for changes in undocumented data averages (^{Feng and Wang, 2010}). In RHTest the homogenization process of the data series was also conducted with settings per quantile (QM), in cases where the data were not homogeneous (^{Wang et al., 2010}). Once applied the quality analysis to data series, an analysis of descriptive statistics was performed to determine abnormalities in thermal and precipitation variables using the normal climatic values according to the criteria established by WMO in a 50 years period.

Determination of climatic indices. For this document 8 out of 27 climate indices established by the “expert team on climate change detection and indices" (ETCCDI) were selected (^{Vazquez, 2010}; ^{Zhang et al., 2011}). The selected climate indices were those related to temperature and precipitation (Table 1). The indices were obtained from the RClimDex module developed by the Climate Research Department from the Meteorological Service of Canada and implemented in R version 3.0.1 (^{Zhang and Feng, 2004}). The RClimDex provides data on monthly or yearly basis, temperature thresholds were as follows; 25 °C for maximum daily temperature and 20 °C as higher daily minimum temperature threshold. After obtaining the indices, it proceeded to perform a correlation analysis with the slope values of the trends of each index for the 16 weather stations analyzed, since it was considered that these are values of the same type for all variables, therefore were not standardized. To evaluate the tendency of these climate indices in the state of Chiapas, the values of each index for all weather stations were averaged, obtaining annual values. With these values were developed trend charts for the 8 indices using a four years moving average.

Table 1 Climate indices analyzed for the state of Chiapas in the period 1960-2009. 

Results and discussion

The results from the number of stations showing positive and negative trends with significance level are shown in Table 2. As can be seen, from 16 stations analyzed 63% resulted positive significant to extreme maximum temperature, 50% for diurnal temperature range, 44% for lower maximum temperature, 38% for tropical nights, 31% for higher temperature minimum, 25% for extreme minimum temperature and only 6% for summer days and consecutive days without rain respectively. Finally the consecutive days without rain or CDD index, presented one weather station with significant positive trend, which represents 6% of all analyzed stations. However, the days without rain or with rain <1 mm in the state of Chiapas are on average 63 days, meaning that most of the year (302 days) there were rains above > 1 mm in the period under study.

Table 2 Number of weather stations showing positive trend, negative trend, significant positive trend and significant negative trend. 

Also in Table 2, it is observed that the number of stations with significant positive trend is higher than the significant negative trend, of the 16 weather stations analyzed and 8 indices, 42 combinations showed significant positive trend, while 18 had significant negative trend. Climate indices that showed the highest number of weather stations with a significant trend correspond to the maximum extreme temperature or TXx with 10 stations, diurnal temperature range or DTR with 8 stations and the lowest maximum temperature TXn with 7 stations.

These results are similar to those reported by Peralta- Hernández et al. (2009) in southeastern Mexico, showing significant trends in increasing the maximum temperature and frequency of hot days. On the other hand, the results also agree with those found by Zarazúa et al. (2014), who mention a significant increase in the frequency of warmer night index in the last decades for the coastal plains of the Gulf of Mexico which includes the states of Veracruz, Tabasco and Campeche. Regarding climate tropical nights and summer days indices, in the period under study was observed that in the state of Chiapas there are summer days almost all year (98% of the year), while tropical nights occur less (77% of year).

Furthermore, the results found in this research are consistent with those reported by ^{Yu and Li (2015)}, as these authors demonstrated significant positive trends in most of the stations analyzed taking into account the summer days and tropical night indices, thus TNn and TNx during the period 1960-2011 in northeast China. However, it is noteworthy that the number of stations and the scale they used were higher than the ones studied in this work, which can represent significant differences by including information from a greater number of weather stations.

The mean and standard deviation obtained from the different climate indices are presented in Table 4.

Table 4 Average values and standard deviation of climatic indices analyzed. 

Figure 3 shows the trend of each climate index analyzed, using the four year moving average. It can be seen that the indexes showed different behaviors, for example for summer days index presented in Figure 7a showing an increasing trend with marked oscillations, presenting more pronounced peak values both increase and decrease, generally observing annual variations with a tendency to increase from 2008. However, the difference between 1960 and 2009 was 3 days, being 1976 the year with the lowest average with 354 days and 2009 the highest with 363 days. Figure 3b presents the frequency trend of summer days, it can be observed a slight trend towards increase, the results showed that the highest average was recorded in 1994 with 310 days, while the lowest was in 1968 with 250 days. The results found in this study agree with those reported by ^{Wang et al. (2013)}, these authors found significant positive trend for summer days and tropical nights on 53 and 46% of the stations analyzed in the period 1960-2010, in the arid region of China, reporting that the magnitude of the regional trend was 1.96 and 1.59 days/decade.

Figure 3 Trend of climate indices, a) summer days; b) tropical nights; c) maximum extreme temperature; d) higher minimum temperature; e) lower maximum temperature; f) minimum temperature extreme; g) diurnal temperature range; and h) consecutive days without rain 

The tendency of extreme high temperature or TXx is shown in Figure 3c, the highest value was present in 3 different years, 1986, 2003 and 2009 with values of 38 °C and the lowest in 1999 with 34 °C, a difference of 4 °C. Figure 3c shows the oscillations in average temperature but with a clear trend towards increase. The trend of the highest minimum temperature (TNx) shown in Figure 3d, where it can be seen low variability in time series averages analyzed, however, there is a slight tendency to decrease. The highest temperature was recorded in 1968 with 23 °C and the lowest in 2009 with 17 °C, with a difference of 6 °C between the lowest and highest average. These results agree with those reported by ^{Qiong et al. (2014)} when analyzing the stations located in the region of the Yangtze River in China during 1962 to 2011, where it was reported that most of the stations showed significant positive trend (66%) and a lower percentage with significant negative trend (15%) for maximum extreme temperature.

Regarding the tendency of the lowest maximum temperature (TXn) shown in Figure 3e. It can be seen that in the period 1960-1980, this parameter did not present a defined trend; however, from 1980 to 2000, there was a slight increase. In the decade from 2000 to 2009 the trend was explained similarly to 1960-1980, decreasing. The highest temperature was recorded in 1987 and 1988 with 23 °C, while the lowest was in 1997 with 18 °C. These results do not agree with those reported by ^{Qiong et al. (2014)}, as these authors found significant tendencies towards increase for TXn in 94% of the stations analyzed in the region of the Yangtze River in China during 1962-2011. The trend of extreme minimum temperature or TNn is shown in Figure 3f; variability in the data with marked oscillations along the time series analyzed was observed, peaks can be seen both temperature increase and decrease thereof.

The highest temperature occurred in the years 1969, 1990 and 1994 with a value of 11 °C, the lowest temperature was observed in 2008 with 5 °C. Overall there is a trend towards decline from 1960 to 2009.

The diurnal temperature range (DTR) trend along the time series analyzed is shown in Figure 3f. Over the years little variability was observed, but with a clear tendency to increase, the highest temperature was recorded in 2007 and 2009 with a temperature of 15 °C, while the lowest was in 1961, 1963 and 1999 with temperature of 12 °C. These results do not agree with those reported by ^{Wang et al. (2013)}, since this author mentions that about 75% of the stations analyzed showed significant trend to decrease the diurnal temperature range in the arid zone from China. However, mentions the importance of examining the data series seasonally instead of annually.

Finally, Figure 3h shows the precipitation trend (CDD), the records showed that the lowest average for CDD was in 2000 with 44 days, while the highest was in 2003 with 80 days. There were oscillations both to increase and decrease of CDD, no marked trends were observed, only slight decrease, however this result agrees with that reported by ^{Gimesi (2009)}, reporting that precipitation showed a slight downward trend globally in Eastern Europe; however, did not observed a clear trend when analyzing the data series on annual or seasonal basis. Also, mentions that is worth noting how precipitation decreased in June, but at the same time increased in July in the period of study.

Figure 4 presents the 4 climate indices that obtained the highest number of stations with significant positive trend in the state of Chiapas. Unfortunately in the Eastern region of the state of Chiapas there was not enough information from the weather stations, so that the calculation of the indices was not performed. However, considering the stations taken into account in this study, it was observed that a large number of stations showed a tendency to increase in temperature, as in the case of the maximum temperature extreme and tropical nights (Figure 4a and 4b). Although on a smaller scale, the lower maximum temperature and diurnal temperature range indices had about 40% of the stations with significant positive trend. In general, there were no marked trends by region, since the stations with significant positive trend were present across the state.

Figure 4 Spatial distribution of weather stations presenting significant positive trend: a) extreme maximum temperature; b) Diurnal temperature range; c) lower maximum temperature; and d) tropical nights 

Conclusions

Most of the indices analyzed showed tendency to increase, being the most important extreme maximum temperature, diurnal temperature range, lower maximum temperature and tropical nights. In mapping the results showed no regional trends. All indices based on temperature showed similar and consistent behavior patterns with overall tendency towards warming and these findings are consistent with previous studies in the region and elsewhere.