Six decades of climate sampling

In the overall approach, accounting all the stations without geographic categorization, the values for maximum and minimum temperature remained relatively constant through the 1961-2020 six-decade period (Figure 2). On the other hand, regarding precipitation levels, the decades of 1981-1990 and 2010-2020 showed the wettest conditions (1981-1990 being significantly wetter than 1961-1980, and 1991-2000), as shown in Figure 2.

Figure 2 One-way ANOVA for maximum and minimum temperature, and monthly precipitation at the mid-low Nazas-Aguanaval watershed for the period of 1961-2020 divided in six decades. Confidence intervals are shown adjacent to mean values. Distinct letters account for different statistical groups from a Tukey test. 

Climate change for the geographical categories within the study area

There were certain evidences of climate change within elevation categories for temperature. For the maximum temperature, the period of 1991-2000 was identified as the coldest period in the high elevation category (Figure 3). It was also evident that the minimum temperature was significantly variable for the low elevation category, showing a warming trend for the period of 2000-2020. On the other hand, precipitation showed significant variability in all categories (Figure 3). Moreover, for the low and mid elevation categories the wettest period was 1981-1990; while for the high elevation category, the period of 1991-2000 was the driest (Figure 3).

Figure 3 One-way ANOVA for maximum and minimum temperature, and monthly precipitation at the mid-low Nazas-Aguanaval watershed for the period of 1961-2020 divided in six decades. Analyses were performed among the three groups in each category of A) elevation, B) longitude and C) latitude to evaluate how such physical factors have shaped the climate at a regional scale. Confidence intervals are shown adjacent to mean values. Distinct letters account for different statistical groups from a Tukey test (P < 0.05).  

For the longitude categories, maximum temperature remained relatively stable through the six- decade period. While for minimum temperature, the east category showed a warming trend during 2000-2020. In regards to precipitation, all categories showed 1981-1990 as a wet period; however, 2000-2020 was also notably wet for the west category (Figure 3).

Maximum temperature remained relatively stable through the six-decade period for all latitude categories. While south latitude was the only category that showed significant variability, displaying a warm period during 2000-2010. For the case of precipitation, all categories displayed 1981-1990 as a wet period, however, 2010-2020 was remarkably wet for the north latitude category (Figure 3).

Aside from the overall and geographical categories variability, it was possible to identify decadal periods in which there was meteorological variability for certain categories (Table 2). Based on the maximum temperature behavior (which only displayed overall variability within the high elevation category, see Figure 3), during 1961-1970 and 2001-2010 the low elevation category was significantly warmer (than the other categories). While, during 1971-1980, the west category displayed warmer conditions (than the other categories). Regarding latitude, the center category displayed cooler conditions (than the other categories) during 1961-1970.

Regarding the minimum temperature, the low elevation category displayed significant variability and showed the most constant warming trend (among all geographical categories); the low elevation category was significantly warmer than the other elevation categories during 1991-2020 (Figure 3, Table 3). A similar warming trend was displayed for the east category (which also displayed significant variability within the time periods) during 1981-2020. Comparably, the west category displayed warm conditions during 2011-2020 (Figure 3, Table 3). On the other hand, the south latitude category displayed variability (within the time periods), and the north category displayed cooler conditions (than the other latitude categories) during 1971-1980 and 1991-2020.

Even though precipitation had significant variability, both with and without geographical categories (for all cases), there were specific periods when conditions were either drier or warmer for a given geographical category (Table 2). For elevation categories, the decades of 1971-1980 (only for low elevation), 2001-2010 (in which mid elevation was wetter than high elevation), and 2011-2020 (for all three categories) were wetter (Figure 3, Table 3). On the other hand, 1991 2000 was the driest decade for all three categories. In the case of longitude, the east category displayed the driest conditions (among the longitude categories) during 1971-2010, while during 2001-2020 the west category was the wettest. Regarding latitude, all three categories displayed similar precipitation values, except for 2011 2020, when the north category displayed wetter conditions (than in other latitude categories).

Six decades of climate sampling

In Mexico and worldwide (^{Pavia et al. 2009}, ^{Trenberth et al. 2015}, ^{Stahle et al. 2016}, ^{Broecker 2017}), there was a generalized warming after 1970. For Mexico in general, the increase in temperature has occurred rather in the maximum temperature than in the minimum according to ^{Pavia et al. (2009)}; it is known that for the Northern State of Zacatecas (^{Brito-Castillo et al. 2009}) and the Comarca Lagunera influence area (^{Inzunza-López et al. 2011}) there have been decadal cooling trends for both the maximum and minimum temperatures during the last half of the 20th century. In contrast to all these references, given the analysis of the 26 meteorological stations considered herein, both maximum and minimum mean temperatures remained relatively stable (for 1961-2020), as shown in Figure 2. Regarding the precipitation component, warm North American deserts have displayed episodes of drought and prolonged dry conditions (for 1942-1975, and as of 1999 to present), alternated with relatively wet periods (during 1976-1998), probably due to the intercontinental circulation patterns, such as the Pacific Decadal Oscillation (with a return period of ten years) and the El Niño Southern Oscillation (with a year-to-year return period) (^{Hereford et al. 2006}, ^{Stahle et al. 2016}). Such extreme conditions in turn affect regions of Mexico and the Southwestern United States, and originate the so-called ‘North American Monsoon’, which is a summer-autumn intense rain season (^{Adams and Comrie 1997}, ^{Stahle et al. 2016}, ^{Pascale et al. 2017}). However, since the study area is located in the middle of the two Sierra Madre Mountain complexes in Central Northern Mexico, it is geographically ‘isolated’ and the weather conditions (for example, rain) that hit the coastal areas switch to different intensity after the circulation reaches the mountains, in a phenomenon called orographic rain or shadow (^{Scheinvar et al. 2020}, ^{Zavala et al. 2020}). Since the study area has been classified as semi-arid and arid with a precipitation range of 250-450 mm per year (^{Sánchez-Salas et al. 2014}), those areas that are located westwards could get a larger amount of rain than the eastern or central longitude sites; in contrast, southern sites are likely to get drier (than center or north latitudes) in the future (if the trend observed for the period of 2001-2020 continues, see Table 3).

Climate change for geographical categories within the study area

The reports made by ^{Brito-Castillo et al. (2009)}, ^{Pavia et al. (2009)} and ^{Inzunza-López et al. (2011)} in which cooling trends were found for the period of 1950-2000 at some portions of Durango and Zacatecas did not consider specific time periods for the temperature change, instead they presented absolute temperature change rates (it is also noticeable that these studies did not analyze data after the year 2000). Moreover, there was a warming trend country-wide for the same period (^{Pavia et al. 2009}, ^{Cuervo-Robayo et al. 2020}); contrariwise, we found some evidences of regional cooling during certain decadal periods. The high elevation category displayed a cooling period during 1991-2000 (in maximum temperature), but after said period, the maximum temperature increased (in this category) and it was even significantly warmer by the 2011-2020 decade (Figure 3). The west longitude category displayed cooler conditions (regarding minimum temperature) than the other longitude categories during 1981-2010 (Table 3). Besides the foregoing, there were notably warming trends for the low elevation (during 1981-2020), the east longitude and the north latitude categories (both during 1991-2020) as evidenced by the minimum temperature. Moreover, Torreón, Coahuila, a city within the categories of low elevation, east longitude and north latitude displayed the highest ‘urban heat island’ rate for Mexico during the period of 1950-2000 according to mean temperature increases (^{Jáuregui 2005}); for the case of Torreón, both maximum and minimum temperatures showed warming trends for the period of 1961-2020 (data not shown here). Another reference for local warming can be traced to the Mapimí Biosphere Reserve (located within the northeastern part of the study area), which displayed an increase of days per year with temperatures above 40 °C from 2004-2009 in comparison to 1995-2003, as reported in the Climate Change Action Plan for such Protected Area (^{CONANP 2014}). In this regard, the average maximum temperature did not show significant variability for both the east and north categories, while there were warming trends identified (for minimum temperature during 2001-2020) in said geographical categories in this study. Moreover, a generalized warming could be expected in the future for high elevation and southern sites in terms of maximum temperature (if the trends found here continue). Conversely, in western northern sites with high or low elevations, cooling trends might be prevailing (Table 3), as also reported by ^{Inzunza-López et al. (2011)}.

Since the focus of this study is a semi-arid region within the desert, processes related to the status of ecosystem degradation include drought, primary production and carrying capacity, soil degradation, water resources and social factors (^{Hillel and Rosenzweig 2002}). Future climate projections for the Nazas-Aguanaval region indicate that in the mid to long term, the area is likely to get warmer (^{López-Santos and Martínez-Santiago 2015}). However, in a climatic reconstruction for the study region (not shown here) there were two municipalities (Lerdo and Nazas in Durango) that displayed stable conditions for both the maximum and minimum temperatures, probably because of the buffer effect that running water bodies and covering vegetation provide, since these have conserved wetland forests (^{Lobell and Bonfils 2008}). The downscaling of data into a categorization of geographical features, such as latitude, longitude and elevation, made possible to point out different trends in temperature and precipitation changes within this unique semi-arid zone in terms of biodiversity (^{Sánchez et al. 2014}, ^{Barrows et al. 2016}). In this sense, a fine scale spatial resolution is needed to adequately capture regional modes in complex orographic basin-based settings. Objectively identified localities can be employed not only in tracking regional climate signatures, but also to improve the understanding of the mechanisms behind regional climate variability and climate change (^{Abatzoglou et al. 2009}).