Thermal Regime Characteristics of Alpine Springs in the Marginal Periglacial Environment of the Southern Carpathians

Abstract

Mountain watersheds play a crucial role in sustaining freshwater resources, yet they are highly vulnerable to climate change. In this study, we investigated the summer water temperature of 35 alpine springs in the highest part of the Retezat Mountains, Southern Carpathians, between 2020 and 2023. During the four-year monitoring period, water temperatures across all springs ranged from 1.2 °C to 10.5 °C. Springs emerging from rock glaciers had the lowest average temperature (2.37 °C), while those on cirque and valley floors were the warmest (6.20 °C), followed closely by springs from meadow-covered slopes (6.20 °C) and those from scree and talus slopes (4.70 °C). However, only four springs recorded summer temperatures below 2 °C, suggesting a direct interaction with ground ice. The majority of springs exhibited temperatures between 2 and 4 °C, exceeding conventional thresholds for permafrost presence. This challenges the applicability of traditional thermal indicators in marginal periglacial environments, where reduced ground ice content within rock glaciers and talus slopes can lead to spring water temperatures ranging from 2 °C to 4 °C during summer. Additionally, cold springs emerging from rock glaciers displayed minimal daily and seasonal temperature fluctuations, highlighting their thermal stability and decoupling from atmospheric conditions. These findings underscore the critical role of rock glaciers in maintaining alpine spring temperatures and acting as refugia for cold-adapted organisms. As climate change accelerates permafrost degradation, these ecosystems face increasing threats, with potential consequences for biodiversity and hydrological stability. This study emphasizes the need for long-term monitoring and expanded investigations into water chemistry and discharge dynamics to improve our understanding of high-altitude hydrological systems. Furthermore, it provides valuable insights for the sustainable management of water resources in Retezat National Park, advocating for conservation strategies to mitigate the impacts of climate change on mountain hydrology and biodiversity.

1. Introduction

Often referred to as the “water towers” of the world, mountain watersheds serve as a vital source of freshwater for downstream lowland regions [1,2]. However, these fragile environments are highly vulnerable to climate change, posing a serious threat to global water security [3]. The rising air temperatures endanger cryospheric water resources in high mountain regions, including snow, glaciers, and permafrost [4,5]. As glaciers continue to retreat, interest in ground ice stored in periglacial landforms, such as rock glaciers, as an alternative water source has grown [6,7]. On a global scale, [8] estimated that intact rock glaciers contain 62.02 ± 12.40 Gt of ice. However, ongoing climate change, characterized by rising temperatures and accelerated permafrost thaw [9], threatens the stability of periglacial areas. As permafrost degrades, the loss of water retention capacity jeopardizes freshwater availability and increases the risk of slope destabilization, erosion, and shifts in landscape dynamics [10]. Additionally, the thawing of permafrost in mountainous regions may also trigger acid rock drainage, with potentially harmful consequences for water quality and aquatic ecosystems [11]. These changes have widespread consequences for local ecosystems and human populations, highlighting the need for sustainable management practices to safeguard water resources and mitigate climate-induced environmental impacts [12].

Understanding the spatial distribution of mountain permafrost is crucial due to its role as a freshwater source and its impact on hydro-geomorphic processes in steep slope environments [13]. The temperature of spring water in late summer provides useful evidence of permafrost occurrence, and many researchers have used it for mapping permafrost distribution across catchments [14]. Previous studies have shown that spring water affected by permafrost tends to exhibit lower temperatures compared to unaffected sources [15,16]. Based on research conducted in the Swiss Alps [17,18,19], a thermal classification has been developed to assess the likely presence of permafrost (

Table 1. Thermal classification of spring water temperature according to literature [17,18,19].

Water Temperature (°C)	Interpretation
>2	Absence of permafrost
1–2	Possible presence of permafrost
<1	Probable presence of permafrost

). These thresholds can vary slightly due to local geology, topography, and groundwater flow [14].

The most compelling evidence of mountain permafrost is the presence of rock glaciers [20]. These striking alpine landforms consist of a mixture of unconsolidated rock debris and interstitial ice and are characterized by a distinctive ridge-and-furrow surface topography [21]. Due to the internal deformation of the frozen sediment, these landforms are creeping downslope [22]. Depending on their movement and internal ice presence, rock glaciers are categorized as active, inactive, or relict [20]. Active rock glaciers contain a substantial amount of ice and slowly creep downslope, ranging from a few millimeters to several meters per year [23]. Inactive rock glaciers still retain ice but are no longer moving, while relict rock glaciers have lost both ice and mobility. The term “intact rock glaciers” refers to both active and inactive rock glaciers and is primarily used when information on their kinematics is unavailable [24].

In recent decades, several studies have identified permafrost in the Southern Carpathians through geophysical investigations, rock glacier movement monitoring, and ground surface temperature measurements [25,26,27,28]. Findings show that permafrost in this mountainous region exists under marginal conditions and is patchily distributed, primarily influenced by local topo-climatic factors [28]. This complexity makes permafrost distribution challenging to assess, requiring additional data to develop accurate permafrost distribution models. Such models would be invaluable for spatial planners, national park managers, and tourism stakeholders. Measuring spring water temperature in late summer is a reliable and efficient method for mapping permafrost occurrence [29]. However, this approach has primarily been applied in regions with discontinuous permafrost and abundant ground ice, such as the Swiss Alps. In the Carpathians and other mountainous regions, where permafrost is confined to certain small areas, the traditional spring water temperature thresholds may be inadequate. A recent study on the kinematics of rock glaciers in the Retezat Mountains revealed that most intact rock glaciers exhibit movement in their upper sections, while their fronts remain relatively stable [27]. This suggests that springs influenced by permafrost may be warmed by non-permafrost areas as they flow downslope for several hundred meters before reaching the rock glacier front. In such cases, we hypothesize that spring water temperatures may slightly exceed 2 °C at the rock glacier front. According to [29], permafrost can still exist in seemingly relict rock glaciers, with spring water temperatures reaching up to 3.7 °C. These rock glaciers, which appear relict from a morphological perspective but still contain patches of ice, are known as “pseudo-relict rock glaciers” [30]. The Retezat Mountains host the highest concentration of pseudo-relict rock glaciers in the Romanian Carpathians [24], where permafrost has been documented above 2000 m, typically in areas with large boulders, reduced solar radiation, and a strong cooling effect from the coarse, blocky material [26].

Springs in marginal periglacial regions are essential indicators of permafrost presence and provide valuable insights into the hydrological stability of alpine environments [29]. In terms of sustainability, studying these thermal thresholds is crucial for predicting the effects of climate change on freshwater resources. Cold rocky landforms also act as critical climate refugia for both terrestrial and aquatic biodiversity in mountain ecosystems [31]. Numerous microbes, plants, and animals depend on cold habitats, such as rock glaciers, for survival [32]. Therefore, our study aims to analyze the distribution of cold springs in late summer within a marginal periglacial environment of the Carpathian Mountains and to connect these observations with the distribution of permafrost. To achieve this goal, we employed field measurements using a hand-held digital thermometer and monitored spring water temperatures with data loggers. To minimize uncertainties, we conducted spring water temperature measurements during the July–September period between 2020 and 2023, while accounting for intra-annual variability.

2. Materials and Methods

2.1. Study Area

Spring water temperature measurements were conducted in the central region of the Retezat Mountains (45°20′ N, 22°23′ E). Located in the western sector of the Southern Carpathians (Figure 1), this mountain range is one of the highest in Romania, with Peleaga Peak reaching an elevation of 2509 m. The study area exhibits a transitional continental-temperate climate, with alpine climate characteristics becoming more pronounced above 1800 m. The mean annual air temperature (MAAT) isotherm of 0 °C is situated at an elevation of approximately 2050 m, and at this elevation, the annual precipitation averages around 1000 mm/year [24]. Above 2000 m, the snow cover lasts for more than 180 days per year, typically from November to June [33]. At elevations above 2000 m, the 1991-2020 climate norm was 0.8 °C warmer than the 1961-1990 reference period.

Glaciers are no longer found in the Retezat Mountains; however, during the Last Glacial Maximum (24–19 ka BP), they extended to elevations between 1000 and 1400 m in this region [34]. The Retezat Mountains contain the highest number of rock glaciers in the Romanian Carpathians (94) and exhibit the greatest density of these features in the region (0.52 landforms/km²) [24]. Approximately 10% (9.3 km²) of the total area above 1850 m in the Retezat Mountains is likely underlain by permafrost [28]. Above 1850 m, the geomorphological landscape is characterized by a variety of glacial and periglacial landforms, including glacial cirques, glacial valleys, moraines, rock glaciers, talus slopes, scree slopes, and block streams. In the Retezat Mountains, frozen ground is predominantly seasonal, while permafrost is limited to small areas with optimal conditions for its preservation [26].

Geologically, the Retezat Mountains are primarily composed of granodiorites along with epimetamorphic crystalline schists [35]. The treeline is situated at approximately 1700 m, above which subalpine shrubs and alpine herbaceous vegetation dominate.

Figure 1. (a) Overview map showing the location of the Retezat Mountains within the Southern Carpathians and Romania. (b) Spatial distribution and classification of the investigated springs, along with the location of rock glaciers, overlaid on a hillshade derived from the LAKI II DEM [36].

Figure 1. (a) Overview map showing the location of the Retezat Mountains within the Southern Carpathians and Romania. (b) Spatial distribution and classification of the investigated springs, along with the location of rock glaciers, overlaid on a hillshade derived from the LAKI II DEM [36].

2.2. Field Activities

We investigated 35 mountain springs in the central and highest part of the Retezat Mountains. In the alpine domain, the main sources of spring water are underground water, rain, snow, and permafrost. Based on their position and source, the springs were classified into four categories: (1) springs emerging from rock glaciers, (2) springs originating from scree and talus slopes, (3) springs from slopes covered with meadows, and (4) springs located in cirque or valley floors (Figure 2). Springs associated with rock glaciers are those emerging at the front or along the lateral margins of these landforms. Rock glaciers in the Retezat Mountains are composed of coarse debris and are at least 15–20 m thick [26]. Another common category of springs in the study area includes those seeping from other periglacial landforms, such as scree and talus slopes. While scree slopes are primarily composed of coarse debris are considerably thinner than rock glaciers, talus slopes in the Southern Carpathians typically consist of a mixture of fine and coarse sediments and are generally less than 10 m thick [37]. Springs without a clearly identifiable landform at their source but located on steep slopes (>15°) covered by meadows were classified separately. Additionally, springs emerging from the relatively flat floors of glacial cirques and valleys were considered a distinct category.

To support the thermal monitoring of the springs, we conducted both expeditionary and continuous temperature measurements between July and September, when the terrain is free of snow (Figure 3). Most of the springs were investigated using a handheld digital thermometer during one to three visits per season between 2020 and 2023. Temperature readings were typically taken between 2020 and 2023 using a Testo 110 (Testo AG, Lenzkirch, Germany) instrument with an immersion probe equipped with an NTC temperature sensor, offering an accuracy of ±0.2 °C and a resolution of 0.1 °C. Water temperature measurements were taken by shielding the spring from direct sunlight and ensuring that the probes did not come into contact with sediments, rocks, or vegetation. Each temperature measurement was recorded for at least 10 min, or until the temperature stabilized. Field surveys were carried out to examine the spatial variability of spring water characteristics across different years. During each of the expeditionary measurement campaigns, air temperatures were consistently above 0 °C, even in the nights prior to the surveys. In addition, the campaigns were scheduled for at least two days after precipitation events.

Continuous temperature monitoring of spring water was conducted at seven sites using data loggers. Continuous temperature monitoring allows for the identification of fluctuations and the detection of changes in the thermal behavior of springs. The data loggers are devices equipped with temperature sensors that can be submerged in water to measure and record temperature data at selected intervals. Due to their technical features, these compact devices enable the monitoring and storage of thermal data over an extended period. Data collection began in the summer of 2021 for four springs and in 2022 for two sites, and at one site, monitoring was limited to the 2023 warm season. Temperature was recorded at a 1 h sampling frequency using digital data loggers (Tinytag Aquatic 2). The logger’s accuracy is specified as ±0.5 °C with a resolution of 0.01 °C [38].

While most measurements were conducted at the exact spring location, for springs emerging from scree areas, measurements were taken some distance away from the source (ranging from meters to hundreds of meters). To determine and record the spatial coordinates of the selected sites for thermal measurements of the springs, a Garmin 65S GPS was used, offering location accuracy within ±3 m.

2.3. Data Analysis

The GIS analysis was performed using ArcGIS Pro 3.3 (Esri, Redlands, CA, USA). For the cartographic representations of the Retezat Mountains, we used a 1 m resolution digital elevation model provided by ANCPI through the LAKI MNT dataset [36]. This model was corrected for potential errors in runoff modeling by filling sinks in the topographic surface using the Fill function, and the resulting raster was used for all representations. The statistical analysis and corresponding graphs were performed in Microsoft Excel. Daily precipitation data from the Țarcu meteorological station were obtained from the National Administration of Meteorology.

3. Results

The monitored springs for thermal measurements are located at altitudes ranging from 1692 m to 2220 m in the central part of the Retezat Mountains. The investigated springs were classified as follows: eight springs (RS 1, RS7, RS9, RS18, RS19, RS22, RS25, RS33) originating from rock glaciers, nine springs (RS3, RS4, RS6, RS8, RS11, RS14, RS27, RS28, RS34) emerging from scree and talus slopes, nine springs (RS2, RS12, RS13, RS15, RS16, RS20, RS24, RS26, RS35) having source areas on slopes covered with meadows, and nine springs (RS5, RS10, RS17, RS21, RS23, RS29, RS30, RS31, RS32) located on cirque and valley floors. Springs originating from rock glaciers are found at elevations between 1906 m and 2193 m and exhibit the lowest mean temperature of 2.4 °C, with values ranging from 1.2 °C to 4 °C. Notably, four springs—located in Judele (RS1), Valea Rea (RS18), Pietricelele (RS25), and Pietrele (RS33)—record temperatures below the potential permafrost threshold of 2 °C. The springs emerging from scree and talus slopes have a mean temperature of 4.7 °C, with values ranging from 2.7 °C to 7.7 °C. Only five of these springs have temperatures below 4 °C (RS3, RS4, RS6, RS27, and RS28). All the springs originating from slopes covered with meadows have temperatures above 4 °C, with a mean of 6.2 °C. The same mean temperature of 6.2 °C was recorded for springs located on cirque and valley floors, where temperatures range from 3 °C to 10.5 °C.

Spring temperatures measured during summer 2021 using a portable digital thermometer are shown in Figure 4a. This year was selected for illustration as it represented the period with the greatest number of springs investigated. The data highlighted that the lowest values were recorded in springs originating from rock glaciers, followed by those from scree and talus slopes. In contrast, the highest temperatures were observed in springs from slopes covered with meadows and those located on cirque and valley floors. Springs originating from rock glaciers and slopes covered with meadows indicated more stable thermal conditions, in contrast to the greater variability observed in springs from scree/talus slopes and valley floors. The distribution of temperature values across elevations (Figure 4b) suggests that altitude is not the primary controlling factor; rather, the source of the springs plays a more significant role. While the lowest temperatures are found in springs between 1900 m and 2200 m, 11 springs within this elevation range also exhibit temperatures above 4 °C. Notably, the spring at the lowest elevation (1692 m) has a relatively low temperature (3.4 °C), which is an interesting exception.

Continuous temperature measurements from data loggers monitoring three springs emerging from rock glaciers recorded during the warm seasons between late June 2021 and September 2023 are illustrated in Figure 5a. The minimal temperature fluctuations indicate a high degree of thermal stability, with all three springs maintaining relatively low temperatures across all three summers. For springs originating from rock glaciers, interannual temperature variations remain minimal, staying below 1 °C. While RS33 exhibits a consistently uniform temperature regime, slight variations were recorded at RS1 and RS19, particularly during the 2023 warm season. Figure 5b displays continuous temperature measurements from two springs: one originating from a rock glacier and the other from a scree slope. Monitoring was conducted over two consecutive years during the warm seasons of 2022 and 2023. The RS18 spring, emerging from a rock glacier, maintained a stable temperature regime, with an average of 1.7 °C, a consistent maximum of 2.2 °C in both years, and a minimum of 1.4 °C. In contrast, the RS34 spring, sourced from a scree slope, exhibited significantly greater temperature fluctuations, exceeding 3 °C per season.

A similar pattern of interannual thermal stability in springs originating from rock glaciers is presented in Figure 6. Four such springs exhibited minimal thermal variations when measured with a portable digital thermometer during expeditionary campaigns in the warm seasons of 2020, 2021, 2022, and 2023. The lowest temperatures for all four springs were recorded in the summer of 2022, with RS18 reaching a minimum of 1.2 °C. The highest temperature, 2.8 °C, was recorded in 2021 at the RS7 spring (Figure 6a). Figure 6b presents expeditionary thermal measurements from four springs originating from sources other than rock glaciers. RS17 emerges from cirque and valley floors, while RS12 and RS20 originate primarily from slopes covered with meadows. In contrast, RS34 has its source in a scree and talus slope. All the springs presented in Figure 6b exhibit minimal thermal variations from year to year. However, except for RS17, the recorded temperatures are relatively high for a periglacial environment. As RS17 is located near the front of a rock glacier, its distinct thermal behavior could be attributed to its setting within a permafrost-affected transition zone.

The spatial pattern of spring water temperatures during the warm season (2020–2023) (Figure 7) reveals cold temperatures (<2 °C) at the fronts of four rock glaciers: Judele, Pietrele, Pietricelele, and Valea Rea. In another four cases, springs originating from rock glaciers remain cool (2–4 °C) throughout the warm season. Springs emerging from scree and talus slopes are also generally cool, particularly those located on the northern slope of Custura Bucurei Peak and in the west-facing glacial cirque below Peleaga Peak. In the Valea Rea glacial valley, three springs on the valley floor, as well as one emerging from meadow-covered slopes, are also cool. In contrast, springs in Pietrele Valley, located on the valley floor, have temperatures warmer than 4 °C. Springs near the Galeșu rock glacier are generally warm (>4 °C), with the exception of the spring emerging directly from the Galeșu rock glacier, which remains cool.

Temperature variations of six springs, measured at varying distances downslope from the rock glacier fronts, are depicted in Figure 8. In all cases, the spring temperatures are low at their emergence from the rock glacier front (e.g., 1.2 °C to 2.4 °C) and gradually increase downstream. For instance, RS18 registers a temperature of 1.2 °C at the rock glacier front, rising to 5.4 °C at 100 m downstream. A similar abrupt temperature increase is observed in RS25. The other four springs also show warming with increasing distance, though their temperatures generally remain between 2 °C and 4 °C. The temperature rises at a rate between 0.24 °C and 1 °C within the first 50 m from the front of the rock glaciers. Although spring temperatures at the front of the rock glaciers initially remain below the thermal threshold of 2 °C, they rise significantly over a short distance, surpassing the threshold for permafrost occurrence.

4. Discussion

Analyzing 35 spring water temperature measurements conducted during the warm seasons of the 2020–2023 period, we observed that springs emerging from rock glaciers are significantly colder than others. However, temperatures fall below 2 °C in only four cases (RS1, RS 33, RS18, and RS25), which, according to the classical thresholds [17], suggests that these springs likely originate from ground ice. The remaining four springs emerging from rock glaciers have temperatures between 2 °C and 4 °C, exceeding the traditional thermal threshold for permafrost-derived water.

Previous geophysical and thermal measurements conducted in the Retezat Mountains [27] identified ground ice in the Galeșu rock glacier, although the temperatures of the spring emerging from this rock glacier range between 2 °C and 4 °C. The study by [27] revealed that permafrost is present only in the upper section of the Galeșu rock glacier, while the middle and lower parts no longer contain ground ice. At this site, the distance from the ground ice location to the front of the rock glacier is approximately 200 m, and it is likely that the water temperature increases as it flows downslope, becoming progressively warmer near the front. During multiple field surveys of this landform, the sound of water flowing between the boulders was audible at a depth of just a few meters along most of its length. Similar findings have been reported in the Italian Alps by [29], who observed spring temperatures exceeding 3 °C in rock glaciers containing ground ice. Given that these empirical temperature thresholds were originally established in regions with discontinuous permafrost—where ground ice is widespread across landforms—we suggest that their application be approached with caution in regions with marginal permafrost.

Our experiment, which involved measuring the temperatures of six springs seeping from rock glacier fronts at various downslope distances, demonstrated that water temperature can increase by 0.24 to 1 °C per 50 m. This suggests that as water travels farther from its permafrost source, the thermal conductivity of surrounding materials gradually increases its temperature. Consequently, while the temperature at the rock glacier front may exceed 2 °C, the water may still originate from areas underlain by permafrost. We hypothesize that in regions with isolated permafrost patches, spring temperatures emerging from permafrost may be slightly higher than 2 °C but should not exceed 4 °C if the ground ice source is located a few hundred meters from the rock glacier front (Figure 9). Therefore, in these regions, spring water temperature alone is insufficient for mapping permafrost occurrence, and complementary methods are required.

Previous geophysical measurements also confirmed permafrost occurrence in the Judele, Pietrele, and Pietricelele rock glaciers, where spring water temperatures remain below 2 °C (RS1, RS33, and RS25), supporting the presence of ground ice (Figure 10). Valea Rea, the largest rock glacier in the Southern Carpathians [35], was also investigated, with temperature measurements taken from four springs seeping from this landform. Only one (RS18) of these springs registered a temperature below 2 °C, while the others (RS17, RS20, and RS22) ranged between 2 °C and 4 °C. Previous measurements of the temperature at the bottom of the snow cover (BTS) in late winter, as well as ground surface temperature (GST) monitoring, indicate that permafrost is also present in the upper part of the Valea Rea rock glacier (Figure 10).

A recent study by [28] highlighted that permafrost occurrence in the Southern Carpathians is also associated with talus and scree slopes. Spring water temperature measurements recorded between 2020 and 2023 showed that no springs emerging from these landforms had temperatures below 2 °C. However, in five cases (RS3, RS4, RS6, RS27, and RS28), temperatures ranged between 2 °C and 4 °C. Similarly, three springs located on cirque and valley floors exhibited temperatures within the same range. Without validation data, the presence of permafrost at these sites remains uncertain. Most of the springs originating from slopes covered with meadows recorded consistent high-water temperatures (>4 °C). The analysis of spring water temperatures across different elevations revealed that altitude is not the primary factor influencing water temperature. Instead, the source of the springs plays a more significant role. The coldest springs, which originate from rock glaciers, are found between 1950 and 2150 m in elevation (Figure 9). The key factors that contribute to the cool temperatures of springs originating from rock glaciers include the likelihood of embedded ice, coarse surface clasts, the presence of voids and air circulation, and the absence of soil and vegetation. All these factors create a distinct microclimate within the rock glacier body, characterized by cool temperatures and a significant thermal offset versus air temperature during the warm season.

The continuous monitoring of spring water temperature revealed that the intra-annual and interannual variations in the temperatures of alpine springs are relatively small. Springs emerging from rock glaciers exhibit even lower temperature fluctuations compared to other spring categories. These findings are in accordance with other studies also reporting very small temporal variability from continuous recording of water temperature [15,39]. Minor temperature variations are observed following rain events, where a slight increase may occur. However, within 1 to 3 days after the rain, the temperature regime of the springs stabilizes and returns to normal (Figure 11).

Recent studies [40] have revealed a sharp increase in spring water temperatures in high mountain regions over the past two decades, driven by rising air temperatures. This warming trend is expected to lead to a decline in the biodiversity of cold alpine springs and cause shifts in community composition in the future [41]. This calls for a strong need for long-term records of spring water temperature in alpine environments to better understand how quickly they are warming in different mountainous regions worldwide, as these springs serve as refugia for cold-adapted organisms [31]. Unlike other studies that measured alpine spring water temperatures in only a single season [42], this study employed a longer recording period (2020–2023) to account for variations across multiple years and reduce the influence of outlier years with unusual weather patterns. However, we plan to continue the long-term monitoring of mountain springs and expand our investigations to include other key characteristics, such as water conductivity, water chemistry, and discharge monitoring. According to climate projections, the frequency of droughts in Europe is expected to increase by the end of the century [43], impacting the runoff of mountain streams, which may exhibit intermittent flow regimes [44]. Consequently, more studies on stream runoff are needed in high mountain areas, where subsurface ice is expected to play an increasingly vital role in stream flow, especially as glaciers continue to recede [29]. Based on such findings, National Park administrations could take steps to protect remaining pristine ecosystems and manage aquatic networks in high mountain regions sustainably, in order to prevent further biodiversity loss [45].

5. Conclusions

In this study, we investigated the temperature of 35 alpine springs during the warm season in the highest part of the Retezat Mountains between 2020 and 2023. We found that springs originating from rock glaciers are significantly colder than those from other sources. Only four springs had temperatures low enough (<2 °C) to be considered affected by permafrost. Thirteen springs were cool, with temperatures ranging from 2 to 4 °C, which, based on traditional spring water temperature thresholds, suggests that they do not interact with ground ice. However, our findings indicate that some of these springs originate from intact rock glaciers, highlighting the need for caution when applying classical thresholds for permafrost occurrence, especially in marginal periglacial environments. Cold springs emerging from rock glaciers exhibit lower daily and seasonal temperature fluctuations compared to other springs. These springs, originating from rock glaciers with embedded ice, are more thermally decoupled from atmospheric conditions, although rain episodes can cause a slight temperature increase for 1–3 days. This study highlights the crucial role of rock glaciers in influencing alpine spring temperatures, a key factor as these springs act as refugia for cold-adapted organisms. Furthermore, monitoring spring water temperatures from rock glaciers enhances our understanding of permafrost occurrence in marginal periglacial environments, where its distribution is highly complex. These environments are particularly vulnerable to climate change, as rising temperatures can completely thaw the remaining ground ice in the following decades, threatening slope stability and impacting runoff in mountain streams. Given the novel insights into alpine spring water temperatures in remote high-mountain areas, this research provides a foundation for sustainable water resource management in Retezat National Park, the oldest national park in Romania. Additionally, it emphasizes the importance of long-term monitoring and interdisciplinary approaches to mitigate the risks of permafrost degradation, while calling for more sustainable management measures for streams in high mountain regions to address threats to alpine biodiversity.