Adequate estimates of near‐surface temperature lapse rate (γ_local) are needed to represent air temperature in remote mountain regions with sparse instrumental records such as the mountains of central Asia. To identify the spatial and temporal variations of γ_local in the Tianshan Mountains, long‐term (1961–2011) daily maximum, mean, and minimum temperature (T_max, T_mean, and T_min) data from 17 weather stations and 1 year of temperature logger data were analyzed considering three subregions: northern slopes, Kaidu Basin, and southern slopes. Simple linear regression was performed to identify relationships between elevation and temperature, revealing spatial and seasonal variations in γ_local. The γ_local are higher on the southern slopes than the northern slopes due to topography and regional climate conditions. Seasonally, γ_local are more pronounced higher in the summer than in the winter months. The γ_local are generally higher for T_max than T_mean and T_min. The Kaidu Basin shows similar seasonal variability but with the highest γ_local for T_mean and T_min occurring in the spring. Formation of γ_local patterns is associated with the interactions of climate factors in different subregions. Overall, annual mean γ_local for T_max, T_mean, and T_min in the study's subregions are lower than the standard atmospheric lapse rate (6.5°C km⁻¹), which would therefore be an inadequate choice for representing the near‐surface temperature conditions in this area. Our findings highlight the importance of spatial and temporal variations of γ_local in hydrometeorological research in the data‐sparse Tianshan Mountains.

1 Introduction

Mountain regions, as an important supplier of snow‐ or rain‐fed freshwater to lowlands, are significant hydrological and climatological drivers [De Jong et al., 2005]. Ecosystems and environmental processes in mountain regions are highly sensitive to climate change [Beniston, 2003; Barry, 2008]. The Tianshan Mountains, known as the “water tower of central Asia,” are a large system of mountain ranges located in central Asia and are highly susceptible to climate variations [Sorg et al., 2012; Chen, 2014; Hu et al., 2014]. Climate in this region has been changing from warm‐dry to warm‐wet based on historical recorded data [Shi et al., 2007], and water resources are expected to become more unstable in this semiarid alpine region [Piao et al., 2010; Shen and Chen, 2010]. Water resource assessments and the investigation of other climatically driven Earth surface phenomena such as mountain permafrost require the accurate regional representation of near‐surface air temperature, which is strongly dependent on elevation. However, insufficient measurements and the complex topography pose a challenge for temperature extrapolation and hydrological modeling over mountain regions [Minder et al., 2010; Ayala et al., 2015; Jobst et al., 2016]. Spatial and temporal variabilities of climate drivers (i.e., temperature and precipitation) in Tianshan Mountains are poorly understood owing to the sparsity of recording data.

Constant or spatially uniform temperature lapse rates were typically adopted for extrapolating air temperature from meteorological stations to different elevations [Minder et al., 2010]. Temperature lapse rate (γ) is the rate of temperature change with elevation, also known as the vertical temperature gradient [Whiteman, 2000]. From atmospheric physics, unsaturated air parcel cools at the dry adiabatic lapse rate of 9.8°C km⁻¹ when displaced upward, while the saturated adiabatic lapse rate varies with temperature and local conditions due to the condensation process [Whiteman, 2000; Barry, 2008; Barry and Chorley, 2009]. The environmental temperature lapse rate is the actual temperature decrease with height at a given place and time, which depends on the local vertical profile air temperature [Dodson and Marks, 1997; Barry and Chorley, 2009]. In general, it is often replaced by the standard atmospheric lapse rate in the troposphere (6 or 6.5°C km⁻¹) [Harlow et al., 2004; Barry, 2008; Minder et al., 2010]. The standard atmosphere lapse rate is commonly used for temperature extrapolation due to lack of measurements [J. Wang et al., 2010; Luo et al., 2012]. However, applying a standard lapse rate has its challenges since spatiotemporal patterns of temperature lapse rate can be influenced by diurnal cycles [Pepin et al., 1999; Pepin, 2001; Sheridan et al., 2010], direction of wind (i.e., windward or leeward side differences [Minder et al., 2010]), and general contrasts between slope sites [Rolland, 2003; Tang and Fang, 2006; Kattel et al., 2015]. Many studies suggest that γ_local values are typically higher for maximum temperatures than minimum and generally lower in winter than in summer or spring [Rolland, 2003; Tang and Fang, 2006; Blandford et al., 2008]. It has been found that in air temperature extrapolation, regionally defined near‐surface temperature lapse rates perform better than the constant lapse rate [Harlow et al., 2004; Lundquist and Cayan, 2007; Blandford et al., 2008; Petersen et al., 2013]. Near‐surface temperature lapse rate is the rate of change in temperature with elevation observed typically at 2 m above the surface. It is typically decreased with increasing elevation; however, the opposite effect may occur under certain conditions [Blandford et al., 2008]. It was therefore explored from observed temperature data (recorded at 2 m above the surface) using linear regression. From here forward, we use γ_local (γ_local = −∂T/∂Z) to refer to the locally computed near‐surface temperature lapse rate. Throughout this paper, a positive temperature lapse rate refers to when temperature decreases with elevation and a higher/steeper γ_local refers to a rapid temperature decrease (lower/shallower γ_local refers to slow decrease) with elevation as done in previous studies [Minder et al., 2010; Petersen and Pellicciotti, 2011].

The γ_local is a critical factor for mountain hydrology. Adequate estimate of γ_local is not only important for temperature downscaling [Gardner et al., 2009] but also essential for distinguishing the precipitation form (rain or snow) [Minder et al., 2010]. In addition, γ_local is of great significant for snowmelt runoff and glacier melt change [Chutko and Lamoureux, 2009; Lundquist and Cayan, 2007; Deng et al., 2015]. Previous studies have emphasized the spatial and temporal patterns of γ_local in mainland China, whereas they subjectively divided the whole China into different climate zones or southern and northern parts [Li et al., 2013, 2015]; the patterns of γ_local may not reveal regional details in mountain areas due to irregular distribution and sparse weather stations at higher elevations. At a regional scale, Zhang et al. [2012] modified a monthly degree‐day model for semiarid northwestern China by considering different monthly γ_local based on latitude zones; however, the range of latitude zones is too coarse to adequately capture γ_local variations which vary with time and topography in mountainous regions. Some researchers adopted 6°C km⁻¹ as a regional temperature lapse rate [Luo et al., 2012], while others applied 5°C km⁻¹ [Yang et al., 2014] in the Tianshan Mountains. Furthermore, climate differs substantially between the northern/southern slopes and Kaidu Basin in the Tianshan Mountains [Chen, 2014]. As described previously, there is still a lack of knowledge on possible spatial and seasonal variations of γ_local in the Tianshan Mountains, and in particular, possible surface‐based temperature inversions in winter are rarely discussed.

This study aims to evaluate the spatial and temporal variations of γ_local by separating the Tianshan Mountain into three subregions based on topography and different climate regimes: the north facing slopes (northern slopes), the inner mountain region (Kaidu Basin), and the south facing slopes (southern slopes). Fifty‐one years of climate records and a 1 year record of temperature logger data were used to illustrate the spatial and temporal patterns of γ_local in this area. This study is not only important for hydrometeorological research in Tianshan region; it would also be most useful for water resource research in central Asia.

2 Study Area, Data Sets, and Methods

2.1 Study Area

The Xinjiang Tianshan Mountains (located 41°N–45°N, 80°E–88.5°E) are surrounded by the Taklamakan desert to the south and the Junggar Basin to the north (Figure 1). The study area is characterized by an arid continental climate and elevations ranging from −65 to 6804 m above sea level (asl). The Tianshan Mountains can be divided into three general subregions according to topography and climatic conditions: northern slopes, the Kaidu Basin, and southern slopes. Yining and Zhaosu stations are influenced by westerly winds, and their climatic features are more similar to the northern slopes; they will be used to characterize the climate of the northern slopes. The distribution and summary of stations are shown in Figure 1 and Table 1. Annual mean temperatures on the southern slopes (9.7°C) are higher than on the northern slopes (6.7°C).

**Figure 1**
Open in figure viewer PowerPoint

Study area and spatial distribution of weather stations: (a) location within central Asia, (b) overview of subregions and climate stations over the Tianshan Mountains, (c) location of HOBO logger stations in the Kaidu Basin.

Table 1. Weather Stations Used in This Studya^a H1–H9 indicates the HOBO logger data from field survey. S, N, and Kaidu indicate the southern slopes, northern slopes, and Kaidu Basin, respectively. (Bayinbuluke station was excluded from the analysis in Kaidu Basin due to different data availability than logger data.)

ID	Name	Subregions	Latitude (°N)	Longitude (°E)	Elevation (m asl)	T_mean (°C)	Precipitation (mm)	Start Year–End Year
1	Yanqi	S	42.08	86.31	1055	8.58	76	1961–2011
2	Kuerle	S	41.75	86.01	932	11.78	55	1961–2011
3	Kuche	S	41.72	82.51	1082	11.30	70	1961–2011
4	Luntai	S	41.78	84.11	976	11.20	65	1961–2011
5	Baicheng	S	41.78	81.51	1229	7.94	119	1961–2011
6	Akesu	S	41.17	80.11	1104	10.45	74	1961–2011
7	Baluntai	S	42.73	86.11	1739	7.12	208	1961–2011
8	Kumishen	S	42.23	88.11	922	9.51	53	1961–2011
9	Urumqi	N	43.78	87.32	935	7.22	264	1961–2011
10	Shihezi	N	44.32	86.01	443	7.48	212	1961–2011
11	Wusu	N	44.43	84.4	479	8.14	170	1961–2011
12	Caijiahu	N	44.2	87.30	441	6.19	143	1961–2011
13	Dabancheng	N	43.35	88.12	1104	6.66	70	1961–2011
14	Yining	N	43.95	81.2	663	9.16	278	1961–2011
15	Jinghe	N	44.62	82.51	320	7.92	104	1961–2011
16	Wenquan	N	44.97	81.00	1358	3.98	232	1961–2011
17	Zhaosu	N	43.15	81.01	1851	3.40	509	1961–2011
18	Bayinbuluke	Kaidu	43.03	84.02	2458	−4.25	272	1961–2011
19	H1	Kaidu	42.71	83.93	2428	−2.37	‐	09.2014–08.2015
20	H2	Kaidu	42.89	83.71	2470	−4.52	‐	09.2014–08.2015
21	H3	Kaidu	42.77	84.56	2483	−2.06	‐	09.2014–08.2015
22	H4	Kaidu	42.94	84.17	2525	0.36	‐	09.2014–08.2015
23	H5	Kaidu	42.69	83.69	2663	−2.65	‐	09.2014–08.2015
24	H6	Kaidu	42.92	83.33	2791	−0.90	‐	09.2014–08.2015
25	H7	Kaidu	43.14	85.50	2986	−0.96	‐	09.2014–08.2015
26	H8	Kaidu	43.19	85.51	3427	−2.67	‐	09.2014–08.2015
27	H9	Kaidu	43.22	85.53	3771	−4.35	‐	09.2014–08.2015

^a H1–H9 indicates the HOBO logger data from field survey. S, N, and Kaidu indicate the southern slopes, northern slopes, and Kaidu Basin, respectively. (Bayinbuluke station was excluded from the analysis in Kaidu Basin due to different data availability than logger data.)

The distribution of precipitation also shows important spatial variation (Figure S1 in the supporting information). Annual average precipitation is 90 mm on southern slopes and 220 mm on the northern slopes according to recorded data. The Kaidu Basin drains approximately 18,649 km² and has a mean elevation of 3100 m asl. It is a typical snowmelt and precipitation‐fed basin that stretches over parts of central Tianshan Mountains and mountainous valley, and it can be recognized as an independent mountain climate system. The Tianshan Mountains are the most important water source for downstream rivers, which lead to the Tarim Basin on the southern slope and the Yili Valley and Junggar Basin on the northern slope.

Overall, temperature in the Tianshan Mountains is highly elevation dependent and is characterized by warm summers (southern slopes: 23.2°C and northern slopes: 21.8°C) and cool winters (southern slopes: −6.7°C and northern slopes: −11.3°C) based on average temperature. Precipitation in mountain areas is more than the lower part depending on elevation [Chen, 2014]. Snowfall can also occur in summer at higher elevations (i.e., the snowline at 4000–4100 m asl in Urumqi River basin on the northern slopes of Tianshan Mountains [Zhao et al., 2006]); the precipitation and hydrothermal processes in this region are very complex due to the vertical differentiation of temperature.

2.2 Data

Long‐term time series of daily maximum, mean, and minimum temperatures (T_max, T_mean, and T_min, respectively); precipitation; relative humidity; and wind speed data were obtained from China's Meteorological Administration (CMA). Stations record observation data 4 times a day (Beijing time 02, 08, 14, and 20 h). The T_max and T_min were determined from these four measurements, while T_mean is the average. This data set is composed of 18 climate stations covering the period of 1961–2011 in and around the mountain area (Figure 1b and Table 1). The homogeneity of the meteorological data was assessed by using penalized maximal t tests to detect multiple change points in data series [Wang et al., 2007; Wang, 2008; Wang and Feng, 2013] (RHtestsV4 package, available from the ETCCDI website: http://etccdi.pacificclimate.org/software.shtml). A 1.5°C temperature shift was detected from the Baluntai station in 1995, which could be explained by a change in the station's location. The closest station (Yanqi station) was chosen as a reference, which is assumed has the same climate signal (trend and periodic components). Data from Baluntai were adjusted by quantile matching [X. L. Wang et al., 2010; Vincent et al., 2012].

Since CMA climate stations have poor coverage of the inner mountain Kaidu Basin, field data collection was necessary to improve our estimation of γ_local. Thus, HOBO Pro v2 (U23‐001) temperature/relative humidity data loggers (HOBO) were set up in the Kaidu basin to monitor near‐surface temperature (2 m above surface) and relative humidity between 2428 and 3771 m asl (Figure 1c and Table 1) from September 2014 to August 2015. Monthly and annual data were computed based on daily data. The HOBO data loggers can better represent the monthly temperature pattern in the Tianshan Mountains (Figure S2). The HydroSHEDS void‐filled digital elevation model at 3 arc sec resolution was furthermore used (approximately 90 m [Lehner et al., 2008]; http://www.worldwildlife.org/hydrosheds).

2.3 Exploratory Analysis

Annual temperatures were averaged for T_max, T_mean, and T_min from 51 years of data records. The spatial and temporal variations in γ_local were explored using simple linear regressions of temperature and elevation for each of the three subregions (northern slopes, Kaidu Basin, and southern slopes; Figures 1b and 1c). Loess models [Cleveland and Devlin, 1988] were furthermore used to examine nonlinear altitudinal temperature gradients. We analyzed these data sets separately due to different data availability: 51 years of climate records for the northern slopes and southern slopes and only 1 year of field survey data for the Kaidu Basin, respectively. Furthermore, we explored the relationship between γ_local for T_mean with climate factors. The Pearson's correlation coefficient (r) was used to measure the strength of linear associations between elevation and temperature.

3 Results

3.1 Annual Variations: The Southern Slopes Versus the Northern Slopes

The mean annual γ_local shows very strong variation and in particular substantial differences between different slopes (Figures 2a–2c). The γ_local for T_max, T_mean, and T_min on the southern slopes are 5.2, 4.8, and 3°C km⁻¹ and on the northern slopes are 3.1, 3, and 2.5°C km⁻¹, respectively. The γ_local on the southern slopes are substantially higher than the northern slopes. Additionally, mean annual γ_local for T_max is the highest, followed by mean annual γ_local for T_mean and T_min. Comparatively, mean annual γ_local for T_max, T_mean, and T_min on both sides of Tianshan Mountains are lower than the standard atmospheric lapse rate of 6.5°C km⁻¹. Thus, an extrapolation of air temperature using this standard lapse rate would be expected to produce a substantial bias in regional‐scale estimates.

**Figure 2**
Open in figure viewer PowerPoint

Altitudinal variation in mean annual temperature ((a) T_max, (b) T_mean, and (c) T_min) with their respective linear fit and γ_local for both the southern slopes (circles) and northern slopes (triangles) of the Tianshan Mountains. For comparison, the cross indicates the BYBLK station, located in the Kaidu Basin. Correlations in Figure 2a are southern slopes: r = −0.91, p value < 0.01; northern slopes: r = −0.81, p value < 0.01; in Figure 2b are southern slopes: r = −0.74, p value < 0.05; northern slopes: r = −0.82, p value < 0.01, and in Figure 2c are southern slopes: r = −0.38, p value = 0.35; northern slopes: r = −0.66, p value = 0.05. The p values correspond to the null hypothesis of zero correlation.

The correlation of mean annual T_max with elevation on the southern slopes is stronger than on the northern slopes (r =−0.91 and r =−0.81, respectively). The reverse holds true for mean annual T_mean, whose correlation with elevation on the southern slopes (r =−0.74) is weaker than on and the northern slopes (r =−0.82). However, the relationship between mean annual T_min and elevation is not robust (southern slopes: −0.38 and northern slopes: −0.66; Figures S3 and S4).

3.2 Seasonal Variations: The Southern Slopes Versus the Northern Slopes

The seasonal γ_local over the Tianshan Mountains shows remarkable differences on the different slopes (Figures 3a–3c). Overall, the seasonal cycle of γ_local for T_mean is higher on southern slopes than northern slopes in all seasons. The γ_local for T_max and T_min follow the same distribution except for T_min in autumn, where the average γ_local for T_min on the northern slopes exceeded the value for the southern slopes (2.6 and 2.4°C km⁻¹, respectively). However, γ_local for T_max is higher in summer than γ_local for T_mean and T_min on both sides of Tianshan Mountains. In addition, the lapse rates (γ_local for T_max, T_mean, and T_min) in summer were 8.6, 8.2, and 6.7°C km⁻¹ for the southern slopes and were 7, 6.7, and 5.7°C km⁻¹ for the northern slopes, respectively (Figure 3). Notably, the lapse rates (γ_local for T_max, T_mean, and T_min) in autumn were lower than in spring.

**Figure 3**
Open in figure viewer PowerPoint

Seasonal variations of lapse rates (γ_local for (a) T_max, (b) T_mean, and (c) T_min) calculated by linear regression of temperature and elevation. The bold line indicates the average value (averaged γ_local over the periods of 1961–2011). Seasonal conditions are divided into Spr (spring: March, April, and May), Sum (summer: June, July, and August), Aut (autumn: September, October, and November), and Win (winter: December, January, and February).

The winter season lapse rates on both sides of Tianshan Mountain are more variable than in other seasons (Figure 3). The average ranges of year‐to‐year variation in seasonal average γ_local were 1.8 (standard deviation = 0.44), 2.9 (0.66), 1.9 (0.46), and 4.6 (1.16)°C km⁻¹ for spring, summer, autumn, and winter on the southern slopes and 3.1 (0.55), 2.5 (0.32), 2.3 (0.50), and 3.5 (0.86)°C km⁻¹ on the northern slopes. However, the lapse rates in winter were observed to have either lower or negative values, i.e., surface‐based temperature inversions on both sides of Tianshan Mountains.

3.3 Geographic Variability: Kaidu Basin

The Kaidu Basin differs in elevation, topography, and local climate from the northern and southern slopes. Seasonal variation of γ_local shows slight differences compared to the other sites. Monthly γ_local for T_max between April and October is higher than γ_local for T_mean and T_min in the Kaidu Basin (Figure 4a). The highest γ_local for T_max occurs in September (8.4°C km⁻¹), while γ_local for T_mean and T_min occur in April (6.9 and 5.8°C km⁻¹, respectively). The γ_local for T_mean and T_min have a decreasing trend after April.

**Figure 4**
Open in figure viewer PowerPoint

(a) Variations of γ_local for T_max, T_mean, and T_min in monthly scales (April to October) in the Kaidu Basin (09.2014–08.2015; winter months not included due to nonlinearity). (b) The correlation between elevation and temperature in winter months (average temperature from November to March) and loess smoothers.

A nonlinear trend for lapse rates in the Kaidu Basin was observed in the winter months related to surface‐based temperature inversion (Figures 4b and S5); the temperature (T_max, T_mean, and T_min) generally increases in the valleys and lowlands with increasing elevation and then tends to decrease at higher elevations. For T_max and T_mean, the inflection elevation is around 3000 m asl. However, T_min becomes flat above 2800 m asl. We cannot identify the top of the inversion layer inside Kaidu Basin due to the data sparsity and complexity of inversion process.

4 Discussion

4.1 Mechanisms of Seasonal Variation

The variations of γ_local (γ_local only refer to γ_local for T_mean here and after since γ_local for T_max, T_mean, and T_min have very similar distribution patterns at the monthly scale) show seasonality throughout the year with higher γ_local in the summer months and shallower γ_local in the winter months in all subregions (Figures 3 and 4), which is consistent with previous researches [Rolland, 2003; Blandford et al., 2008; Kirchner et al., 2013; Li et al., 2013]. As seasonally modified radiative and turbulent heat exchanges can significantly affect the temperature structure and the adiabatic process [Barry and Chorley, 2009], we suggest that the seasonal variation of γ_local can be explained by the seasonal variation of solar radiation. Regions in the middle latitudes in the northern hemisphere receive more radiation in summer than in winter, which leads to a similar seasonal temperature variation (higher in summer and lower in winter; Figures 5a and 6a). However, there is regional exception in the Kaidu Basin, where the maximum lapse rate occurs in April (Figure 6a). The instability of atmosphere between snow‐free ground in lowland and snow‐covered mountain slopes could modify lapse rates [Barry and Chorley, 2009]. Melting snow is evident in Kaidu Basin in the spring, which is likely the cause for the higher lapse rates in April.

**Figure 5**
Open in figure viewer PowerPoint

Seasonal cycle of the relation between lapse rate (γ_local for T_mean) and climate factors ((a) temperature, (b) precipitation, (c) relative humidity, and (d) wind speed) on the southern and northern slopes of Tianshan Mountains (1961–2011).

**Figure 6**
Open in figure viewer PowerPoint

Annual cycle of the relationship between γ_local for T_mean with (a) temperature and (b) relative humidity in the Kaidu Basin (09.2014–08.2015).

Surface‐based winter temperature inversions are another possible cause for the seasonal variation of γ_local. Surface‐based temperature inversion is known to weaken the relationship between air temperature and elevation [Marshall et al., 2007; Marshall and Losic, 2011; Cullen and Marshall, 2011; Kattel et al., 2013]. Winter temperature inversions are common phenomenon in Tianshan mountain area. The temperature lapse rates can be reversed from positive to negative due to winter temperature inversions; thus, the γ_local in the winter months are lower than in the summer months.

4.2 Mechanisms of Spatial Variation

Near‐surface temperature lapse rates for all months are higher on the southern slopes than on the northern slopes over Tianshan Mountains, which may result from the difference of slope sites and local climatic conditions. The role of water vapor in the air is essential for the spatial pattern of γ_local. Previous research indicated that shallower γ_local is typically associated with relatively warmer and moister atmospheric conditions [Pepin et al., 1999; Barry and Chorley, 2009; Li et al., 2013; Jobst et al., 2016]. Air parcels cool more slowly in a humid environment than dry climates as it rises because of greater amounts of latent heat that can be released from vapor condensation [Whiteman, 2000; Barry and Chorley, 2009]. Thus, the magnitudes of temperature changes with elevation are reduced. This mechanism can be revealed by the spatial variability of precipitation and humidity, which are higher on the northern slopes than the southern slopes (Figures 5b and 5c). Although an opposite sign was found in temperature that southern slopes are warmer than the northern slopes (Figure 5a), the southern slopes are generally dryer and more limited by precipitation compared to the northern slopes (annual precipitation are 90 and 220 mm for southern and northern slopes, respectively). Therefore, the modification of the lapse rate is still largely due to less latent heat that can be released by the condensation process. Furthermore, the spatial variability of γ_local is also consistent with the pattern of wind speed. Higher wind speed can enhance the air turbulences and the mixing of air masses, which will decrease the temperature gradient [Barry and Chorley, 2009]. Wind speed on the northern slopes is higher than southern slopes (Figure 5d); this mechanism might be another explanation for the lower γ_local on the northern slopes while higher γ_local on the southern slopes.

Data source and length between northern and southern slopes with Kaidu Basin are different; thus, γ_local values from different subregions may not be spatially comparable. Specifically, it is unclear why higher relative humidity occurred in May and June in the Kaidu Basin (Figure 6b). Field measurement shows that there is a vast wetland located in the central Kaidu Basin, which may influence the surrounding climate and introduce additional uncertainty to the relative humidity calculation. We acknowledge that some uncertainties remain. More measurement data are required for detailed analysis.

4.3 Mechanisms of Surface‐Based Temperature Inversion

Surface‐based temperature inversion during the winter was observed for all sites in Tianshan Mountains (southern slopes, Kaidu Basin, and northern slopes; Figures 3-6). There are some possible reasons for this phenomenon in the Tianshan Mountains. Solar radiation will heat ground surface temperature during the day; in turn, the ground will release longwave radiative fluxes in the night, which will cool down surface temperature. Thus, nocturnal radiative cooling at the surface that leads to diurnal change in surface temperature might be the reason for surface‐based temperature inversion [Whiteman et al., 1999; Whiteman, 2000; Barry and Chorley, 2009]. Local topography (i.e., mountain valley) can also modify the atmospheric structure. Cold air drainage above the ground will limit temperature exchange between high elevations and lowlands, which leads to atmosphere subsidence, furthermore reducing cloud formation and enhancing longwave cooling of the surface temperature, and further yields strong inversion [Vihma, 2011]. Specifically, the relationship between temperature and elevation in winter is nonlinear in Kaidu Basin (Figure 4b). Cullen and Marshall [2011] recommended to use piecewise linear models for the regionalization of air temperature where inversion and noninversion periods exist. However, challenges still remain in understanding the surface‐based temperature inversion since inversion depth and intensity always change with season and local climate.

4.4 Implication for Modeling Earth Surface Processes

The findings in this study illustrate that γ_local has a spatiotemporal distribution pattern in Tianshan Mountains and the constant environmental temperature lapse rate (6.5°C km⁻¹) throughout the year cannot represent the variability of the temperature‐elevation relationship in complex terrain areas. The behaviors of γ_local can be quite important for ecosystems and mountain hydrological process research. For instance, higher (lower) γ_local will result in lower (higher) temperature from lower to higher elevations in temperature interpolation; errors therefore not only increase in temperature extrapolation and downscaling [Komatsu et al., 2010; Sheridan et al., 2010] but also result in a reduction on modeled melt when higher lapse rates are used [Gardner et al., 2009; Petersen and Pellicciotti, 2011]. In addition, locally computed temperature lapse rate plays a great role in identifying the form of rain or snow [Singh and Singh, 2001]. Furthermore, the choice of temperature lapse rate has a significant influence on the seasonal discharge variation [Minder et al., 2010; Deng et al., 2015]. It has been observed that the timing of summer melt can shift a full month earlier when the lapse rate is changed from 6.5 to 4°C km⁻¹ [Minder et al., 2010]. In particular, the application of reasonable lapse rate can effectively improve the accuracy of hydrological modeling in data‐sparse alpine catchments [Immerzeel et al., 2014; Jobst et al., 2016]. Additionally, spatial and seasonal changes of γ_local are vital for vegetation‐climate relationships in mountainous areas [Tang and Fang, 2006; Bendix et al., 2009] and modeling the distribution of mountain permafrost based on the significant impact on ground temperatures [Lewkowicz and Bonnaventure, 2011; Gruber, 2012]. Besides, the topography information (valley or slopes) appears to influence the temperature interpolation reliability as well; the interpolation error may be reduced when taking the topographic differences into account in an alpine region [Rolland, 2003].

Spatial and seasonal γ_local are vital to applications such as temperature extrapolation and hydrological modeling in the Tianshan Mountains. Previous study indicated that the runoff change in spring and autumn is highly sensitive to the γ_local in the Kaidu River basin [Deng et al., 2015]. Since snow and glaciers are widespread in the Tianshan Mountains and temperature needs to be extrapolated from the lower elevation due to data sparsity and incomplete coverage, we have suggested to use the seasonal γ_local to reduce uncertainties in related climatic and hydrological research. The constant lapse rate should only be considered as a last resort. This knowledge is also transferable to research in mountainous areas in central Asia where with similar topographic complexity and climatic contrasts.

5 Conclusions

Spatial and temporal variations of γ_local in different slope sites and Kaidu Basin were investigated in the Tianshan Mountains by simple linear regression based on the recorded and field survey data. Seasonally, lapse rates are higher in summer than in winter months in all subregions. Lapse rates for T_max are higher than that for T_mean and T_min. Spatially, lapse rates are higher on the southern slopes than the northern slopes, except for the Kaidu Basin which has the special mountain climate. The standard atmospheric lapse rate (6.5°C km⁻¹) throughout the whole year would therefore mislead results in temperature extrapolation. The spatial and temporal variations of γ_local in Tianshan Mountain are linked to the geographic differences and climate factors and should be taken into account in meteorological and hydrological applications. In addition, it is reasonable to consider surface‐based winter temperature inversions (i.e., nonlinear relationship) in models extrapolating temperature in winter time. The knowledge of this study is useful for hydrometeorology research in the Tianshan Mountain region and regions in central Asia.

Acknowledgments

This study was supported by the Key Project of the Chinese Academy of Sciences (KZZD‐EW‐12‐1). Thanks to the China Scholarship Council for a PhD scholarship (201304910343). The data were provided by China's Meteorological Administration (CMA; http://data.cma.cn/). The fieldwork data in this study have been acquired in collaboration with the Key Laboratory of Agricultural Water Resources, Center for Agricultural Resources Research, CAS. The authors express their sincere thanks to colleagues from the Center for Agricultural Resources Research, CAS, for their kind support during field survey and data collection. Field data may be obtained from Yanjun Shen (e‐mail: yjshen@sjziam.ac.cn). We also wish to acknowledge Manfred Fink, Sven Kralisch, and Miga Magenika Julian for their valuable suggestions. We appreciate the Editor and three anonymous reviewers for their constructive and insightful comments.

Supporting Information

References

Ayala, A., F. Pell icciotti, and J. M. Shea (2015), Modeling 2 m air temperatures over mountain glaciers: Exploring the influence of katabatic cooling and external warming, J. Geophys. Res. Atmos., 120, 3139– 3157, doi:10.1002/2015JD023137.
Wiley Online Library ADS Web of Science®Google Scholar
Barry, R. G. (2008), Mountain Weather and Climate, 3rd ed., Cambridge Univ. Press, Cambridge, U. K.
Crossref Google Scholar
Barry, R. G., and R. J. Chorley (2009), Atmosphere, Weather and Climate, 9th ed., Routledge, Oxon, U. K.
Crossref Google Scholar
Bendix, J., H. Behling, T. Peters, M. Richter, and E. Beck (2009), Functional biodiversity and climate change along an altitudinal gradient in a tropical mountain rainforest, in Tropical Rainforests and Agroforests Under Global Change, edited by T. Tscharntke et al., pp. 239– 268, Springer, Berlin, doi:10.1007/978-3-642-00493-3_11.
Crossref Google Scholar
Beniston, M. (2003), Climatic change in mountain regions: A review of possible impacts, Clim. Change, 59, 5– 31, doi:10.1023/A:1024458411589.
Crossref Web of Science®Google Scholar
Blandford, T. R., K. S. Humes, B. J. Harshburger, B. C. Moore, V. P. Walden, and H. Ye (2008), Seasonal and synoptic variations in near‐surface air temperature lapse rates in a Mountainous Basin, J. Appl. Meteorol. Climatol., 47, 249– 261, doi:10.1175/2007JAMC1565.1.
Crossref ADS Web of Science®Google Scholar
Chen, Y. N. (2014), Water Resources Research in Northwest China, Springer, New York doi:10.1007/978-94-017-8017-9.
Crossref Google Scholar
Chutko, K. J., and S. F. Lamoureux (2009), The influence of low‐level thermal inversions on estimated melt‐season characteristics in the central Canadian Arctic, Int. J. Climatol., 29(2), 259– 268, doi:10.1002/joc.1722.
Wiley Online Library ADS Web of Science®Google Scholar
Cleveland, W. S., and S. J. Devlin (1988), Locally weighted regression: an approach to regression analysis by local fitting, J. Am. Statist. Assoc., 83(403), 597– 610. doi:10.2307/2289282.
Crossref Google Scholar
Cullen, R. M., and S. J. Marshall (2011), Mesoscale temperature patterns in the rocky mountains and foothills region of southern Alberta, Atmos.‐Ocean, 49, 189– 205, doi:10.1080/07055900.2011.592130.
Crossref Web of Science®Google Scholar
De Jong, C., D. Collins, and R. Ranzi (2005), Climate and Hydrology in Mountain Areas, Wiley, West Sussex, England.
Wiley Online Library Web of Science®Google Scholar
Deng, H., Y. Chen, H. Wang, and S. Zhang (2015), Climate change with elevation and its potential impact on water resources in the Tianshan Mountains, central Asia, Global Planet. Change, 135, 28– 37, doi:10.1016/j.gloplacha.2015.09.015.
Crossref ADS Web of Science®Google Scholar
Dodson, R., and D. Marks (1997), Daily air temperature interpolated at high spatial resolution over a large mountainous region, Clim. Res., 8, 1– 20, doi:10.3354/cr008001.
Crossref ADS PubMed Web of Science®Google Scholar
Gardner, A. S., M. J. Sharp, R. M. Koerner, C. Labine, S. Boon, S. J. Marshall, D. O. Burgess, and D. Lewis (2009), Near‐surface temperature lapse rates over Arctic glaciers and their implications for temperature downscaling, J. Clim., 22(16), 4281– 4298, doi:10.1175/2009JCLI2845.1.
Crossref ADS Web of Science®Google Scholar
Gruber, S. (2012), Derivation and analysis of a high‐resolution estimate of global permafrost zonation, Cryosphere, 6(1), 221– 233, doi:10.5194/tc-6-221-2012.
Crossref ADS Web of Science®Google Scholar
Harlow, R. C., E. J. Burke, R. L. Scott, W. J. Shuttleworth, C. M. Brown, and J. R. Petti (2004), Derivation of temperature lapse rates in semi‐arid south‐eastern Arizona, Hydrol. Earth Syst. Sci., 8, 1179– 1185.
Crossref ADS Web of Science®Google Scholar
Hu, Z., C. Zhang, Q. Hu, and H. Qin (2014), Temperature changes in central Asia from 1979 to 2011 based on multiple datasets, J. Clim., 27(3), 1143– 1167, doi:10.1175/JCLI-D-13-00064.1.
Crossref ADS Web of Science®Google Scholar
Immerzeel, W. W., L. Petersen, S. Ragettli, and F. Pellicciotti (2014), The importance of observed gradients of air temperature and precipitation for modeling runoff from a glacierized watershed in the Nepalese Himalayas, Water Resour. Res., 50, 2212– 2226, doi:10.1002/2013WR014506.
Wiley Online Library ADS Web of Science®Google Scholar
Jobst, A. M., D. G. Kingston, N. J. Cullen, and P. Sirguey (2016), Combining thin‐plate spline interpolation with a lapse rate model to produce daily air temperature estimates in a data‐sparse alpine catchment, Int. J. Climatol., doi:10.1002/joc.4699.
Wiley Online Library Google Scholar
Kattel, D. B., T. Yao, K. Yang, L. Tian, G. Yang, and D. Joswiak (2013), Temperature lapse rate in complexmountain terrain on the southern slope of the central Himalayas, Theor. Appl. Climatol., 113, 671– 682, doi:10.1007/s00704-012-0816-6.
Crossref ADS Web of Science®Google Scholar
Kattel, D. B., T. Yao, W. Yang, Y. Gao, and L. Tian (2015), Comparison of temperature lapse rates from the northern to the southern slopes of the Himalayas, Int. J. Climatol., 35, 4431– 4443, doi:10.1002/joc.4297.
Wiley Online Library ADS Web of Science®Google Scholar
Kirchner, M., T. Faus‐Kessler, G. Jakobi, M. Leuchner, L. Ries, H.‐E. Scheel, and P. Suppan (2013), Altitudinal temperature lapse rates in an Alpine valley: Trends and the influence of season and weather patterns, Int. J. Climatol., 33, 539– 555, doi:10.1002/joc.3444.
Wiley Online Library ADS Web of Science®Google Scholar
Komatsu, H., H. Hashimoto, T. Kume, N. Tanaka, N. Yoshifuji, K. Otsuki, M. Suzuki, and T. Kumagai (2010), Modeling seasonal changes in the temperature lapse rate in a northern Thailand mountainous area, J. Appl. Meteorol. Climatol., 49(6), 1233– 1246, doi:10.1175/2010JAMC2297.1.
Crossref ADS Web of Science®Google Scholar
Lehner, B., K. Verdin, and A. Jarvis (2008), New global hydrography derived from spaceborne elevation data, Eos Trans. AGU, 89(10), 93– 94, doi:10.1029/2008EO100001.
Wiley Online Library ADS Google Scholar
Lewkowicz, A. G., and P. P. Bonnaventure (2011), Equivalent elevation: A new method to incorporate variable surface lapse rates into mountain permafrost modelling, Permafrost Periglacial Process., 22, 153– 162, doi:10.1002/ppp.720.
Wiley Online Library Web of Science®Google Scholar
Li, X., L. Wang, D. Chen, K. Yang, B. Xue, and L. Sun (2013), Near‐surface air temperature lapse rates in the mainland China during 1962–2011, J. Geophys. Res. Atmos., 118, 7505– 7515, doi:10.1002/jgrd.50553.
Wiley Online Library ADS Web of Science®Google Scholar
Li, Y., Z. Zeng, L. Zhao, and S. Piao (2015), Spatial patterns of climatological temperature lapse rate in mainland China: A multi‐time scale investigation, J. Geophys. Res. Atmos., 120, 2661– 2675, doi:10.1002/2014JD022978.
Wiley Online Library ADS Web of Science®Google Scholar
Lundquist, J. D., and D. R. Cayan (2007), Surface temperature patterns in complex terrain: Daily variations and long‐term change in the central Sierra Nevada, California, J. Geophys. Res., 112, D11124, doi:10.1029/2006JD007561.
Wiley Online Library ADS Web of Science®Google Scholar
Luo, Y., J. Arnold, P. Allen, and X. Chen (2012), Baseflow simulation using SWAT model in an inland river basin in Tianshan Mountains, northwest China, Hydrol. Earth Syst. Sci., 16, 1259– 1267, doi:10.5194/hess-16-1259-2012.
Crossref ADS Web of Science®Google Scholar
Marshall, S. J., M. J. Sharp, D. O. Burgess, and F. S. Anslow (2007), Near‐surface‐temperature lapse rates on the Prince of Wales Icefield, Ellesmere Island, Canada: Implications for regional downscaling of temperature, Int. J. Climatol., 27, 385– 398, doi:10.1002/joc.1396.
Wiley Online Library ADS Web of Science®Google Scholar
Marshall, S., and M. Losic (2011), Temperature lapse rates in glacierized basins, in Encyclopedia of Snow, Ice and Glaciers, Springer, Netherlands.
Crossref Google Scholar
Minder, J. R., P. W. Mote, and J. D. Lundquist (2010), Surface temperature lapse rates over complex terrain: Lessons from the Cascade Mountains, J. Geophys. Res., 115, D14122, doi:10.1029/2009JD013493.
Wiley Online Library ADS PubMed Web of Science®Google Scholar
Pepin, N. C. (2001), Lapse rate changes in northern England, Theor. Appl. Climatol., 68(1–2), 1– 16, doi:10.1007/s007040170049.
Crossref ADS Web of Science®Google Scholar
Pepin, N. C., D. Benham, and K. Taylor (1999), Modeling lapse rates in the maritime uplands of northern England: Implications for climate change, Arct. Antarct. Alp. Res., 31(2), 151– 164.
Crossref Web of Science®Google Scholar
Petersen, L., and F. Pellicciotti (2011), Spatial and temporal variability of air temperature on a melting glacier: Atmospheric controls, extrapolation methods and their effect on melt modeling, Juncal Norte Glacier, Chile, J. Geophys. Res., 116, D23109, doi:10.1029/2011JD015842.
Wiley Online Library ADS PubMed Web of Science®Google Scholar
Petersen, L., F. Pellicciotti, I. Juszak, M. Carenzo, and B. W. Brock (2013), Suitability of a constant air temperature lapse rate over an Alpine glacier: Testing the Greuell and Böhm model as an alternative, Ann. Glaciol., 54(63), 1– 1, doi:10.3189/2013AoG63A477.
Crossref ADS Web of Science®Google Scholar
Piao, S., P. Ciais, Y. Huang, Z. Shen, S. Peng, J. Li, L. Zhou, H. Liu, Y. Ma, and Y. Ding (2010), The impacts of climate change on water resources and agriculture in China, Nature, 467(7311), 43– 51, doi:10.1038/nature09364.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Rolland, C. (2003), Spatial and seasonal variations of air temperature lapse rates in alpine regions, J. Clim., 16, 1032– 1046, doi:10.1175/1520-0442(2003)016<1032:SASVOA>2.0.CO;2.
Crossref ADS Web of Science®Google Scholar
Shen, Y., and Y. Chen (2010), Global perspective on hydrology, water balance, and water resources management in arid basins, Hydrol. Process., 24, 129– 135, doi:10.1002/hyp.7428.
Wiley Online Library PubMed Web of Science®Google Scholar
Sheridan, P., S. Smith, A. Brown, and S. Vosper (2010), A simple height‐based correction for temperature downscaling in complex terrain, Meteorol. Appl., 17, 329– 339.
Wiley Online Library Web of Science®Google Scholar
Shi, Y., Y. Shen, E. Kang, D. Li, Y. Ding, G. Zhang, and R. Hu (2007), Recent and future climate change in northwest China, Clim. Change, 80(3–4), 379– 393, doi:10.1007/s10584-006-9121-7.
Crossref CAS ADS Web of Science®Google Scholar
P. Singh, and V. P. Singh (2001), Snow and Glaciers Hydrology, Kluwer Acad., Dordrecht, Netherlands.
Google Scholar
Sorg, A., T. Bolch, M. Stoffel, O. Solomina, and M. Beniston (2012), Climate change impacts on glaciers and runoff in Tien Shan (central Asia), Nat. Clim. Change, 2(10), 725– 731.
Crossref Web of Science®Google Scholar
Tang, Z., and J. Fang (2006), Temperature variation along the northern and southern slopes of Mt. Taibai, China, Agric. For. Meteorol., 139, 200– 207, doi:10.1016/j.agrformet.2006.07.001.
Crossref ADS Web of Science®Google Scholar
Vihma, T. (2011), Atmosphere‐snow/ice interactions, in Encyclopedia of Snow, Ice and Glaciers, pp. 66– 75, Springer, Netherlands, doi:10.1007/978-90-481-2642-2_31.
Crossref Google Scholar
Vincent, L. A., X. L. Wang, E. J. Milewska, H. Wan, F. Yang, and V. Swail (2012), A second generation of homogenized Canadian monthly surface air temperature for climate trend analysis, J. Geophys. Res., 117, D18110, doi:10.1029/2012JD017859.
Wiley Online Library ADS Web of Science®Google Scholar
Wang, J., H. Li, and X. Hao (2010), Responses of snowmelt runoff to climatic change in an inland river basin, northwestern China, over the past 50 years, Hydrol. Earth Syst. Sci., 14, 1979– 1987, doi:10.5194/hess-14-1979-2010.
Crossref ADS Web of Science®Google Scholar
Wang, X. (2008), Accounting for autocorrelation in detecting mean shifts in climate data series using the penalized maximal t or F test, J. Appl. Meteorol. Climatol., 47, 2423– 2444, doi:10.1175/2008JAMC1741.1.
Crossref ADS Web of Science®Google Scholar
Wang, X. L., and Y. Feng (2013), RHtestsV4 user manual. [Available at http://etccdi.pacificclimate.org/RHtest/RHtestsV4_UserManual_10Dec2014.pdf.]
Google Scholar
Wang, X. L., Q. H. Wen, and Y. Wu (2007), Penalized maximal t test for detecting undocumented mean shift without trend change, J. Appl. Meteorol. Climatol., 46, 916– 931, doi:10.1175/JAM2504.1.
Crossref ADS Web of Science®Google Scholar
Wang, X. L., H. Chen, Y. Wu, Y. Feng, and Q. Pu (2010), New techniques for detection and adjustment of shifts in daily precipitation data series, J. Appl. Meteorol. Climatol., 49(12), 2416– 2436, doi:10.1175/2010JAMC2376.1.
Crossref CAS ADS Web of Science®Google Scholar
Whiteman, C. D. (2000), Mountain Meteorology: Fundamentals and Applications, pp. 355, Oxford Univ. Press, New York.
Web of Science®Google Scholar
Whiteman, C. D., X. Bian, and S. Zhong (1999), Wintertime evolution of the temperature inversion in the Colorado Plateau basin, J. Appl. Meteorol., 38, 1103– 1117.
Crossref ADS Web of Science®Google Scholar
Yang, T., X. Wang, Z. Yu, V. Krysanova, X. Chen, F. W. Schwartz, and E. A. Sudicky (2014), Climate change and probabilistic scenario of streamflow extremes in an alpine region, J. Geophys. Res. Atmos., 119, 8535– 8551, doi:10.1002/2014JD021824.
Wiley Online Library ADS Web of Science®Google Scholar
Zhang, S., B. Ye, S. Liu, X. Zhang, and S. Hagemann (2012), A modified monthly degree‐day model for evaluating glacier runoff changes in China. Part I: Model development, Hydrol. Process., 26, 1686– 1696, doi:10.1002/hyp.8286.
Wiley Online Library ADS Web of Science®Google Scholar
Zhao, J., S. Zhou, Y. He, Y. Ye, and S. Liu (2006), ESR dating of glacial tills and glaciations in the Urumqi River headwaters, Tianshan Mountains, China, Quat. Int., 144(1), 61– 67, doi:10.1016/j.quaint.2005.05.013.
Crossref ADS Web of Science®Google Scholar

Citing Literature

Volume121, Issue23

16 December 2016

Pages 14,006-14,017

Spatial‐temporal variation of near‐surface temperature lapse rates over the Tianshan Mountains, central Asia

Abstract

1 Introduction