Comparison of the surface energy budget between regions of seasonally frozen ground and permafrost on the Tibetan Plateau

Atmospheric Research 153 (2015) 553–564

Contents lists available at ScienceDirect

Atmospheric Research journal homepage: www.elsevier.com/locate/atmos

Comparison of the surface energy budget between regions of seasonally frozen ground and permafrost on the Tibetan Plateau Lianglei Gu a, Jimin Yao b,⁎, Zeyong Hu a, Lin Zhao b a Nagqu Observatory for High and Cold Climate and Environment/Key Laboratory of Land Surface Process and Climate Change in Cold and Arid Regions, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou 730000, China b Cryosphere Research Station on Qinghai–Xizang Plateau/State Key Laboratory of Cryospheric Science, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou 730000, China

a r t i c l e

i n f o

Article history: Received 4 July 2014 Accepted 13 October 2014 Available online 22 October 2014 Keywords: Seasonally frozen ground Permafrost Energy budget Freezing and thawing progress Monsoon

a b s t r a c t Surface energy budgets were calculated using turbulent flux observation data and meteorological gradient data collected in 2008 from two sites: BJ, located in a seasonally frozen ground region, and Tanggula, located in a permafrost region. In 2008, the energy closure ratios for the BJ and Tanggula sites were 0.74 and 0.73, respectively, using 30-min instantaneous energy flux data but 0.87 and 0.99, respectively, using daily average energy flux data. Therefore, the energy closure status is related to the time scale that is used for the study. The variation in the surface energy budget at the two sites was similar: The sensible heat flux (Hs) was relatively high in spring and reduced in summer but gradually increased in autumn. The latent heat flux (LE) was higher in summer and autumn but lower in winter and spring. Comparably, the starting time for the significant increase in LE occurred earlier at the Tanggula site than that at the BJ site, because the freezing and thawing progress of the active layer of permafrost at Tanggula site significantly affected the Hs and LE distributions, but the freezing and thawing processes of the soil at BJ site did not significantly affect the Hs and LE distributions. The monsoon significantly affected the variation in Hs and LE at both the BJ and Tanggula sites. Regarding the diurnal variation of the energy budget at the two sites, the daily maximum of net radiation (Rn) occurred at approximately 14:00 Beijing Time, and the daily maximum of ground heat flux (G0) was earlier than those of Hs and LE. The albedo and Bowen ratio for the two sites were both low in summer but high in winter. The albedo increased significantly but the Bowen ratio became lower or even negative when the surface was covered with deep snow. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The Tibetan Plateau plays an important role in providing thermodynamic power within the Asian monsoon system. The energy and water cycles of the Tibetan Plateau are important for the formation and development of the Asian monsoon system, and components of the Asian and global energy and water cycles (Ye and Gao, 1979; Ye, 1981; Yanai et al., 1992; Ye and Wu, 1998; Shi et al., 2001; Hsu and Liu, 2003; Sato and ⁎ Corresponding author. Tel.: +86 931 4967718. E-mail address: [email protected] (J. Yao).

http://dx.doi.org/10.1016/j.atmosres.2014.10.012 0169-8095/© 2014 Elsevier B.V. All rights reserved.

Kimura, 2007; Xu et al., 2008; Ma et al., 2008; Cui and Graf, 2009; Duan et al., 2012; Yang et al., 2011, 2014). Permafrost occupies nearly 54.3% of the area of the Tibetan Plateau (Cheng, 2005), where it is an important component of the natural ecosystem of the plateau and a sensitive indicator of climate change (Pavlov, 1994). Seasonally frozen ground covers 1100 × 103 km2 of the Tibetan Plateau (Zhou et al., 2000) and plays an important role on the plateau because most of the important ecological, hydrological, pedological, and biological activities occur within it (Zhao et al., 2004). The differences exist between the seasonally frozen ground and the permafrost because of the effects of different regions,

554

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

Fig. 1. Comparison of the soil temperatures between BJ site (a) and Tanggula site (b).

geological features, hydrothermal features, and physical processes that play important roles in the regional and global climate systems (Yang et al., 2003). By definition, the seasonally frozen ground only freezes for two weeks to several months each year (Roger and Thian, 2011). The seasonally frozen ground is an important portion of hydrologic and climatic variables because of their effects on water supplies, energy exchanges, and climate–cryosphere interactions in the atmospheric boundary layer (Duguay et al., 2005). Permafrost is ground that remains at or below 0 °C for at least two consecutive years, and the near surface formed an active layer along with the transition of seasons (Woo, 2012). The freezing and thawing processes of the active layer of the permafrost affect the migration and distribution of the unfrozen water in the frozen soil, thereby affecting the hydrothermal conduction properties and heat flux of the soil. Additionally, the permafrost influenced the infiltration of rain and snowmelt and affected surface runoff and evaporation. Such processes are important influences on the surface energy budget (Zhao et al., 2000). The monitoring of the energy flux change in the permafrost region is meaningful for the study of frozen soils, the interactive feedback of matter, and the energy exchange between the cryosphere and other major spheres as well as for sensitivity analyses (Ramos et al., 2007). The surface energy, mass, and momentum are changing with climatic warming and may have direct and/or indirect feedbacks to the climate. The energy and water budgets for the regions of seasonally frozen ground and permafrost on the Tibetan Plateau are extremely important and strongly coupled with hydrology and climate as well as with other biological and physical processes related to ecosystem structure and function (Gu et al., 2005; Ma et al., 2009; Duan et al., 2012; Yang et al., 2014). The surface of the Tibetan Plateau absorbs a large amount of solar radiation energy and undergoes dramatic seasonal changes in surface heat and water fluxes that greatly influence regional and global climates. Several atmosphere– land interaction experiments have been conducted on the Tibetan Plateau in recent years (Ma et al., 2003, 2009; Tanaka et al., 2001, 2003). Those scientific experiments have resulted in much progress regarding the understanding of the surface energy and water budget, regional evaporative fraction, the seasonal variability of soil moisture distributions, atmospheric

chemistry, and climatic change (Ma and Tsukamoto, 2002; Hirose et al., 2002; Tanaka et al., 2001, 2003; Li et al., 2007; Yao et al., 2008, 2011; Yu et al., 2008; Cong et al., 2009; Ma et al., 2009, 2012; Zheng et al., 2010; Xue et al., 2013; Ma et al., 2014). However, those studies have been limited to the investigation of differences in the surface energy budget between the seasonally frozen ground region and the permafrost region of the Tibetan Plateau. To understand the mechanisms underlying the surface energy budget processes characterizing the seasonally frozen ground and permafrost regions on the Tibetan Plateau, the present study used the eddy covariance method to calculate the turbulent flux at two sites in 2008: one on seasonally frozen ground (BJ site) and one on permafrost (Tanggula site). The Bowen ratio and mean diurnal variation (MDV) methods were used to fill data gaps. The characteristics and differences in the surface energy budgets at the two sites were analyzed with the goal of contributing to the description and the mechanistic understanding of the land–atmosphere interactions occurring on the Tibetan Plateau. 2. Sites and data This study used two monitoring sites, one located in the Nagqu Valley region (BJ site) and the other located at Tanggula Pass (Tanggula site). The BJ site (91°54′E; 31°22′N; and 4509 m above sea level (a.s.l.)) is located in the alpine steppe of the Nagqu Valley within a plateau subfrigid semi-humid climate zone. The area of the observation site is 8000 m2, and the ground is flat, with open surrounding terrain. The surface is primarily covered with sandy soil, accompanied by a sparse distribution of fine stones. An uneven growth of alpine steppe, with a height of 10–20 cm, occurs during the summer. According to the soil temperature data for the BJ site, which is shown in Fig. 1(a), this site is located in the region of seasonally frozen ground. The soil was frozen from January to February, and it thawed from the ground surface beginning at the end of March. The soil froze from the ground surface beginning in early November, and the maximum freezing depth of the soil was approximately 1.5 m. The Tanggula site (91°52′E; 33°04′N; and 5100 m a.s.l.) is adjacent to the Qinghai–Tibet highway and located on a

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

555

Fig. 2. The monitoring apparatus at BJ site (a is the eddy covariance system, and b is the 10-m meteorological gradient tower) and Tanggula site (c is the eddy covariance system, and d is the 10-m meteorological gradient tower).

gentle slope facing the southwestern Tanggula Pass area. The observation site is flat, with open surrounding terrain. The vegetation is alpine meadow and is distributed in clusters less than 10 cm high. According to the soil temperature data for the active layer at the Tanggula site, which is shown in Fig. 1(b), this site is located within the permafrost region. The soil was frozen from January to March. The active layer thawed from the ground surface beginning at the end of April, and the maximum depth of thawing in the active layer was beyond 3 m in September. The active layer began upward freezing from the permafrost table in early October. After approximately 10 more days, downward freezing began from the ground surface and progressed rapidly. These changes were followed by bidirectional freezing processes. Thereafter, the soil froze again in November and December. The data used in this study were primarily obtained from an eddy covariance system and a 10-m meteorological gradient towers at the BJ and Tanggula sites from January 1, 2008, to December 31, 2008 (in this study, all the times were Beijing Time, which is GMT plus 8 h). Fig. 2 showed the eddy covariance system and the 10-m meteorological gradient towers at the BJ and Tanggula sites. The BJ eddy covariance system was located 20 m above the ground on a 25-m tower.

This system primarily included a 3-D ultrasonic anemometer– thermometer (DAT-600, Kaijo Corp., Japan) and an open-path infrared gas analyzer (LI-7500, Li-Cor, Inc., Lincoln, NE, USA) that were used to measure the 3-D wind velocity, virtual temperature, water vapor density, and CO2 density with a 10-Hz sampling frequency. These data were collected using a data logger (CR5000, Campbell Scientific, Logan, UT, USA). The Tanggula eddy covariance system was located 3 m above the ground on a tripod. This system primarily included a 3-D ultrasonic anemometer–thermometer (CSAT3, Campbell Scientific) and an open-path infrared gas analyzer (LI-7500) that were used to measure the 3-D wind velocity, virtual temperature, water vapor density, and CO2 density with a 10-Hz sampling frequency. A CR5000 data collector was again used to log the data. Table 1 provides detailed information regarding the instruments composing the eddy covariance system. In addition to the turbulence measurements, the upward and downward short-wave and long-wave radiation, soil heat flux, soil moisture content, soil temperature, and snow cover, among other variables, were measured using the 10-m meteorological gradient tower at both sites. Table 2 provides detailed information regarding the instruments located on these towers.

556

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

Table 1 Specification of observation instruments of the eddy covariance system at BJ site and Tanggula site.

BJ

Tanggula

Observation item

Instruments

Manufacturer

Accuracy

Height/depth

3-D wind velocity Virtual temperature Water vapor density CO2 density 3-D wind velocity Virtual temperature Water vapor density CO2 density Pressure

DAT-600 DAT-600 LI-7500 LI-7500 CSAT3 CSAT3 LI-7500 LI-7500 Li-7500

KAIJO KAIJO LI-COR LI-COR Campbell Campbell LI-COR LI-COR LI-COR

±0.5 cm/s 0.025 °C ±0.03 ppm ±0.03 ppm ±0.4 cm/s − ±0.01 ppm ±0.01 ppm ±1.0 hPa

20 m 20 m 20 m 20 m 3m 3m 3m 3m 1m

(J m−3 K−1). The average volume heat capacity is defined as follows:

3. Processing methods 3.1. Net radiation (Rn) The net radiation (Rn) was calculated using data from four radiation components according to the following equation: Rn ¼ Sd−Su þ Ld−Lu

ð1Þ

where Su is the upward short-wave radiation (W m−2), Sd is the downward short-wave radiation (W m−2), Lu is the upward long-wave radiation (W m−2), and Ld is the downward long-wave radiation (W m−2).

8 C ¼ C dry þ ρliq cliq θ5 cm > > > > < C ¼ 0:90  106 J m−3 K−1 dry 3 −3 −1 > > ρliq ¼ 1:00  10 kg m K > > : 3 −1 cliq ¼ 4:18  10 J kg

ð3Þ

where Cdry is the volume heat capacity (J m−3 K−1) of dry soil, and θ5 cm is the soil volumetric water content (%) at a depth of 5 cm. 3.3. Sensible heat flux (Hs) and latent heat flux (LE)

3.2. Ground heat flux (G0) The ground heat flux (G0) was calculated according to the following equation (Tanaka et al., 2003):
 ∂T ∂T ∂T G0 ≈C 0:01  0 cm þ 0:06  5 cm þ 0:03  10 cm þ G10 ∂t ∂t ∂t

cm

ð2Þ where G10 cm is the observed soil heat flux (W m−2) at a depth of 10 cm and C is the averaged volume heat capacity

The sensible heat flux (Hs) and latent heat flux (LE) were calculated via the eddy covariance method every 30 min using the observation data from the eddy covariance system. However, for long-term continuous observation using the eddy covariance system, varying degrees of missing eddy covariance data might exist because of unfavorable weather conditions, the switching of cards, or an insufficient power supply. For missing eddy covariance observation data, the Hs and LE could be calculated by the Bowen ratio method when data from the 10-m meteorological tower were available; when the tower data were also missing, the MDV method could be used.

Table 2 Specification of observation instruments of the 10-m meteorological gradient towers at BJ site and Tanggula site.

BJ

Tanggula

Observation item

Instruments

Manufacturer

Accuracy

Height/depth

Short wave radiation Long wave radiation Soil heat flux Soil temperature Soil moisture content Air temperature Relative humidity Precipitation Snow depth Short wave radiation Long wave radiation Soil heat flux Soil temperature Soil moisture content Air temperature Relative humidity Precipitation Snow depth

CM21 CM21 EKO TS-301 (Pt100) Trime EZ TS-801 (Pt100) HMP-45D NOAH-II SR-50 CM3 CM3 HFP01 107 CS616 HMP-45C HMP-45C T200-B SR-50

Kipp & Zonen Kipp & Zonen MF-81 Okazaki IMKO Okazaki Vaisala ETI Campbell Kipp & Zonen Kipp & Zonen HUKSEFLUX Campbell Campbell Vaisala Vaisala Geonor Campbell

±10% ±10% ±5% ±3% ±1% ±3% ±2% ±0.1 mm ±1 cm ±10% ±10% ±5% ±2.5% ±2.5% ±2% ±2% ±0.1 mm ±1 cm

1.5 m 1.5 m 10, 20 cm 0, 4, 10, 20, 40 cm 4, 20 cm 1, 8.2 m 1, 8.2 m 1m 3m 2m 2m 5, 10, 20 cm 2, 5, 10, 20, 40 cm 5, 10, 20 cm 2, 5, 10 m 2, 5, 10 m 1m 2m

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

557

3.3.1. Eddy covariance method The eddy covariance method used the 3-D wind velocity, virtual temperature, and water vapor density to calculate the Hs and LE using the following equations:

where Cp is the specific heat at a constant pressure (J kg−1 K−1), Lv is the LE of vaporization (W m−2), and Δt and Δq are the temperature (°C) and water vapor pressure (kPa) gradients, respectively.

Hs ¼ ρC p θ0 w0

ð4Þ

LE ¼ ρLV w0 q0

ð5Þ

3.3.3. Mean diurnal variation (MDV) method The mean diurnal variation (MDV) method is an interpolation technique that is based on the temporal auto-correlation of the fluxes (Falge et al., 2001). A missing observation is replaced by the mean of valid values measured on adjacent days at the same time (the same half an hour or with a buffer of ±1 h). In general, a window length no longer than 2 weeks is recommended because for longer periods, nonlinear dependence on environmental variables could introduce considerable uncertainty and large errors (Falge et al., 2001).

6

Lv ¼ 2:5  10 −2323  t

ð6Þ

where ρ is the air density (kg m−3), Cp is the specific heat capacity at a constant pressure (J kg−1 K−1), θ is the potential temperature (K), w is the vertical wind velocity (m s−1), Lv is the latent heat of vaporization (W m−2), q is the specific humidity (kg kg−1), and t is the air temperature (°C). The quality control of the eddy covariance data primarily included an inspection of the raw eddy covariance data, an atmospheric stability test during the observational periods, and a test of atmospheric turbulence development. First, the raw eddy covariance data were strictly screened and were excluded when 1) the sensor signal was abnormal or 2) the missing data exceeded 5% of the 30-min raw data. In addition, the calculation of eddy flux requires a series of corrections, including double coordinate rotations (Kaimal and Finnigan, 1994), spectral correction in the high frequency range (Massman, 2000), spectral correction in the low frequency range (Finnigan et al., 2003), correction of the buoyancy flux (Liu et al., 2001), and the Webb, Pearman, and Leuning (WPL) correction (Webb et al., 1980), among others. These corrections, especially the spectral corrections in the high and low frequency ranges and the WPL correction, often increase the LE flux by approximately 10% (Oncley et al., 2007). 3.3.2. Bowen ratio method Bowen (1926) proposed the Bowen ratio energy balance method (i.e., the BREB method or the Bowen ratio method) based on the surface energy balance equation: for a given surface, the ratio between the energy assigned to Hs and the energy assigned to LE is relatively constant. The most basic assumption is that the air momentum diffusion coefficient (thermal diffusivity) and water vapor turbulent diffusion coefficient are equal. The following formulas express these relationships: Hs ¼

ðRn−G0 Þβ 1þβ

ð7Þ

LE ¼

Rn−G0 1þβ

ð8Þ

where Rn is the net radiation (W m−2), G0 is the ground heat flux (W m−2), and β is the Bowen ratio, which can be written as follows:

β¼

C p ðt 1 −t 2 Þ CpΔt ¼ Lv ðq1 −q2 Þ LvΔq

ð9Þ

4. Results and discussion 4.1. Energy closure ratio (CR) The status of energy closure is an important criterion used to evaluate the quality of flux data (Foken, 2008a,b; Aubinet et al., 2012). Wilson et al. (2002) analyzed observational data at 50 global FLUXNET stations, and their results indicated a frequent lack of energy closure, with an energy imbalance between 10% and 30% at the different observation sites. Many experiments concerning energy balance closure have been reported in the last 30 years: the sum of the Hs and LE is usually less than the surface available energy, which is the difference between the Rn and the G0 (Foken et al., 2006; Foken, 2008a,b; Hendricks-Franssen et al., 2010; Barr et al., 2012; Leuning et al., 2012; Mauder et al., 2013; Stoy et al., 2013). Several reasons for the energy balance closure problem have been discussed in an overview paper by Culf et al. (2004): 1) measurement errors (in particular, the eddy covariance method might underestimate the turbulent flux), 2) because of different balance layers and scales of diverse measuring methods (the energy storage in the canopy and the soil was often discussed), and 3) the inability of the eddy covariance method to capture the eddies at large time scales because of the heterogeneity of the land surface. Foken et al. (2006) suggested that a loss at low frequency was also an important reason for the energy balance closure problem. In addition, previous studies using instantaneous observation data collected on the Tibetan Plateau during the summer and autumn have indicated that turbulent flux accounts for approximately 80% (or less) of the surface available energy (Tanaka et al., 2001, 2003; Yao et al., 2008, 2011). The energy closure ratio (CR) is defined as follows:

CR ¼

Hs þ LE Rn−G0

ð10Þ

when CR closer to 1.00 indicates a better energy closure status. Fig. 3 shows the energy closure status in 2008 using instantaneous flux data for the BJ and Tanggula sites, where the CR was 0.74 and 0.73, respectively. The energy balance closure, especially for the Tanggula site, improved when the CR was estimated based on the daily average energy flux data, and the CR at the BJ and Tanggula sites was 0.87 and 0.99,

558

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

1200

BJ

800

Hs+LE(W/m2)

Hs+LE(W/m2)

1200

400

0

-400 -400

Tanggula

800

400

0 Hs+LE = 0.7441*(Rn-G0) R² = 0.6818

0

400 Rn-G0(W/m2)

800

1200

-400 -400

Hs+LE = 0.7282*(Rn-G0) R² = 0.6599

0

400 Rn-G0(W/m2)

800

1200

Fig. 3. Energy closure status using 30-min instantaneous flux data at BJ site and Tanggula site in 2008.

respectively (see Fig. 4). Therefore, energy closure status is also related to the study's time scale. The monthly CR was relatively low at the BJ site in November and at the Tanggula site in December, with values of −0.52 and 0.43, respectively, mainly because of the thick snow cover during those months (see Fig. 5). The albedo of the snow was quite high, causing the Rn and Hs to be low or even negative, but the evaporation and sublimation of the snow were high because of wind and direct radiation. Therefore, the energy budget did not attain closure, and the CR was poor. 4.2. Seasonal variation in fluxes Fig. 6 shows the variation in the daily average of the surface energy budget at the BJ and Tanggula sites in 2008. The energy budget variation demonstrated certain similarities but primarily differences between the two sites. The Rn and G0 were high in summer and autumn but low in winter and spring at both sites. When the surface was covered with the thick snow during January to early March and late October to December at the BJ site and during mid-January to late February and mid-October to December at the Tanggula site (see Fig. 5), the Rn rapidly declined to low, or even became negative values, and the albedo became high. The annual averages for the G0 at the BJ and Tanggula sites were −0.5 W/m2 and 0.4 W/m2, respectively. Therefore, the overall annual effect was the transport of heat energy outward by the soil to the

atmosphere at the BJ site and the absorption of heat energy by the soil at the Tanggula site. As shown in Table 3, the monthly average for the G0 at the Tanggula site was higher than that at the BJ site from April to September. In addition, the active layer of permafrost at the Tanggula site absorbed more heat energy for thaw the deeper soil. During January to March and October to December, the monthly average for the G0 at the Tanggula site was lower than that at the BJ site because the surface of the Tanggula site was covered by more snow than was present at the BJ site (see Fig. 5). Additionally, more energy was required to melt the snow at the Tanggula site, which reduced the energy transfer into the soil. The absolute values of Rn and G0 for both sites were higher than those previously reported for arid areas, grasslands, forests, and the Arctic (Eugster et al., 2000; Beringer et al., 2005; Leuning et al., 2005; Kalthoff et al., 2006). This difference most likely occurred because the elevations at the BJ and Tanggula sites were relatively high and the surface vegetation of the two sites was sparse and short; therefore, solar radiation was not blocked by the surface vegetation. The seasonally frozen ground and permafrost act as a buffer for global climate warming because the freezing and thawing processes of the seasonally frozen ground and the active layers of permafrost require considerable heat (Eugster et al., 2000). At both sites, the Hs was relatively high in spring, reduced during summer, and gradually increased in autumn; the LE showed opposite seasonal trends. During January to February,

Fig. 4. Energy closure status using daily average flux data at BJ site and Tanggula site in 2008.

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

559

Fig. 5. The daily maximum snow depth at BJ site and Tanggula site in 2008.

the Hs and LE values were relatively low because the Rn was relatively low. Because of frozen soil and rare precipitation (see Fig. 7), that is, because the soil contained little unfrozen water and the vegetation coverage was sparse, the Rn was primarily converted to Hs, which was higher than the LE at both sites during January to February. Beginning in March, the Hs slightly increased as the Rn gradually increased, but the variation trend for the LE was unclear, and the enhancement of the Hs was significantly greater than that of the LE. The LE began to increase significantly and the Hs to decline at the Tanggula site in late April and at the BJ site in mid-May. After mid-May, the monsoon season increased the precipitation and water vapor density. Simultaneously, the seasonally frozen ground at the BJ site completely thawed, and the active layer of permafrost at the Tanggula site gradually reached its greatest thawed depth

Fig. 6. The daily averages of the surface energy fluxes at BJ site and Tanggula site in 2008 (A1/B1: the start time of soil thawing, A2/B2: the start time of monsoon, A3/B3: the end time of monsoon, A4/B4: the start time of soil freezing).

(see Fig. 1), which increased the soil moisture content (see Fig. 8). Furthermore, the growth of vegetation flourished, and the evapotranspiration from soil increased, which caused more rapid increases in the LE, as the Hs continued to decline. Yang et al. (2003) indicated that the surface soil not only transports a significant amount of moisture through evapotranspiration but also transfers heat energy into the atmosphere; therefore, the surface soil inhibits increases in the soil temperature and causes the land–atmosphere temperature difference to decline, thereby reducing the Hs. During June and September, the Hs was low, and the LE predominated at both sites. During this period, the Rn was primarily converted into LE, and this process continued until the end of September at the BJ site and early October at the Tanggula site. Subsequently, the Hs began to increase again, and the LE began to decline. The Hs was once again higher than the LE, dominating the conversion of the Rn. Generally, the Hs is higher and the LE is lower during winter; however, the opposite trend was observed at both sites in the winter of 2008, namely, that the LE was relatively high and the Hs was relatively low. This pattern occurred because relatively deep snow covered the surface at both sites during this period (see Fig. 5), resulting in low Rn values. The melting of the snow caused the LE to increase and exceed the Hs. Studies on the surface energy budget in the high-latitude permafrost regions of the Northern Hemisphere have found a trend for the Hs to decline during summer. This pattern most likely occurred because the thawing process in the active layer during summer consumes considerable energy, causing the energy of surface warming to decline; therefore, the land–atmosphere temperature difference and the Hs decline (Eugster et al., 2000). Similar characteristics were observed at other sites on the Tibetan Plateau. Studies at the Amdo, MS3478, and Haibei sites have shown that the LE plays a dominant role during summer and autumn (Tanaka et al., 2001, 2003; Ma et al., 2003; Gu et al., 2005). Thus, a certain similarity exists regarding the intra-annual variation in surface energy flux at the individual sites on the Tibetan Plateau. However, because of the different types of underlying frozen soil examined in the present study, the seasonal variation in the turbulent energy flux at the two sites also differed regarding certain features. At the BJ site, the seasonally frozen ground began to thaw from the surface downward at the end of March (Fig. 6 A1). However, the LE had not begun to increase because the precipitation and unfrozen water were sparse during this period. According to the precipitation data, the rainy season

560

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

Table 3 The monthly and annual averages of the surface energy fluxes at BJ site and Tanggula site in 2008 (unit: W m−2). BJ

Jan. Feb. Mar. Apr. May. Jun. Jul. Aug. Sep. Oct. Nov. Dec. Avg.

Tanggula

Rn

G0

Hs

LE

Rn

G0

Hs

LE

13.04 40.09 67.06 96.54 120.44 139.81 127.10 123.63 99.28 66.20 −15.79 19.35 74.73

−5.13 −2.72 0.69 3.33 5.86 5.45 4.15 0.76 −0.92 −3.91 −8.49 −5.01 −0.50

16.84 38.64 63.09 89.82 62.36 34.98 28.38 28.51 28.41 31.56 −5.04 −5.03 34.38

2.98 6.06 3.42 4.28 33.35 75.00 76.74 77.29 57.32 33.95 8.81 31.94 34.26

14.76 31.00 75.99 86.70 124.56 150.91 139.67 133.69 107.12 31.43 −16.63 −20.12 71.59

−8.92 −7.09 0.16 3.94 10.30 10.98 8.79 5.16 3.04 −5.42 −8.23 −7.88 0.40

21.28 19.36 45.29 54.04 49.31 49.77 34.24 34.58 25.30 4.07 −17.21 −22.62 24.78

9.65 23.38 16.17 19.29 58.83 92.50 102.31 102.52 77.67 33.68 8.61 17.37 46.83

began on May 12 (Fig. 6 A2), when the LE value began to significantly increase. The precipitation was concentrated in the rainy season, the total precipitation during the rainy season was 512.00 mm, accounting for 86.70% of the total annual precipitation. During this period, the surface water content increased significantly, and the predominance of the LE was more obvious. On September 30 (Fig. 6 A3), the rainy season at the BJ site ended, and the dominant role of the LE also weakened; furthermore, the Hs began to simultaneously increase. Early in November (Fig. 6 A4) at the BJ site, the seasonally frozen ground began to freeze, but the Hs and LE variation trends did not significantly change. Therefore, the monsoon was primarily responsible for the seasonal change in the dominant roles of the Hs and LE at the BJ site, and the soil freezing and thawing processes did not significantly affect the Hs and LE variations. At the end of April (Fig. 6 B1), the active layer of permafrost at the Tanggula site began to thaw from the surface downward. However, the significant increase in the LE at the Tanggula site began at the end of April, before the beginning of the rainy season at this site on May 19 (Fig. 6 B2). Therefore, the thawing of the active layer and the monsoon both affect the surface energy budget, causing rapid increases in the LE. The total precipitation during the rainy season period was 319.67 mm at the Tanggula site, accounting for 78.88% of the total annual precipitation. That precipitation considerably increased the surface water content and caused the obvious LE dominance. The active layer thawed to its greatest depth in mid-September.

In early October (Fig. 6 B3), the active layer began to freeze, and the LE dominance simultaneously began to weaken. The rainy season ended on October 14 (Fig. 6 B4), causing a more rapid LE decrease. Therefore, the thawing and freezing processes in the active layer of the permafrost and the monsoon both significantly affected the Hs and LE at the Tanggula site. In addition, the annual average Hs at the BJ site was greater than that at the Tanggula site because of the considerable land– atmosphere temperature difference at the BJ site. The annual average LE at the Tanggula site was greater than that at the BJ site, and this difference was related to the relatively large soil moisture content at the Tanggula site. 4.3. Diurnal variation in fluxes Typical sunny days were selected during each of the four seasons in 2008 at the BJ and Tanggula sites to analyze the diurnal variation in the surface energy fluxes. The days selected for spring, summer, autumn, and winter were April 21, July 17, October 26, and January 16 for the BJ site and April 21, August 8, October 5, and January 8 for the Tanggula site. As shown in Figs. 9 and 10, the amplitude of the daily variation in the Rn and the G0 was largest in summer and smallest in winter. The daily maximum Rn at both sites occurred at approximately 14:00 h, which is midday locally. The daily maximum G0 was earlier than those of the Hs and LE. The variations of the Hs and LE showed seasonally alternating

Fig. 7. The variations of precipitation at BJ site and Tanggula site in 2008.

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

561

Fig. 8. The variations of soil moisture content at BJ site and Tanggula site in 2008.

characteristics, as described in Section 4.2. During winter and spring, the magnitude of the diurnal variation in Hs was greater than that in LE; however, the opposite pattern occurred during summer. As Fig. 10 shows, the diurnal variation in the surface energy fluxes at the Tanggula site on October 5, 2008, clearly reflected the daily cyclic soil freezing and thawing processes. In the morning as the sun rose, the Rn rapidly increased and changed from negative to positive at 08:30 h, and the energy was primarily used to thaw the soil. Additionally, the soil moisture content increased, and the LE rapidly rose until the surface water decreased to a low level; however, no significant changes in Hs and G0 were observed at the beginning of the thawing process. Thereafter, the Hs and G0 began to increase with the surface temperature. After 18:30 h, the Rn became negative, and the Hs and G0 rapidly decreased to low levels. The LE slowly decreased because of weak water evaporation, the soil began to freeze, and the soil moisture content declined.

4.4. Albedo Albedo is of particular importance in the land surface energy balance and in the Earth's radiation balance that dictates the rate of heating of the land surface under different environmental conditions (Strugnell and Lucht, 2001). The albedo at both the BJ and Tanggula sites was low in summer but high in winter and increased significantly when the surface was snow covered. The annual average albedo was 0.31 at the BJ site and 0.38 at the Tanggula site. The monthly average albedo was higher at the Tanggula site than at the BJ site (see Table 4) because the snow cover duration and thickness were greater at the Tanggula site than at the BJ site. Furthermore, the vegetative growth in summer was lusher at the BJ site than at the Tanggula site, which caused the surface roughness at the Tanggula site to be less than that at the BJ site. All these factors contributed to the higher albedo at the Tanggula site.

Fig. 9. The diurnal variation of the surface energy fluxes at BJ site, on January 16, April 21, July 17, and October 26 in 2008.

562

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

Fig. 10. The diurnal variation of the surface energy fluxes at Tanggula site, on January 8, April 21, August 8, and October 5 in 2008.

The maximum monthly average albedo occurred in November at both sites (see Table 4) because there was heavy snowfall at the BJ site in late October and at the Tanggula site in mid-October that resulted in a thick snow cover at both sites in November. The daily maximum snow depth exceeded 20 cm at the Tanggula site and 10 cm at the BJ site. Therefore, the albedo was quite high at both sites in November. 4.5. Bowen ratio The Bowen ratio is defined as the ratio of the Hs to the LE (Sverdrup, 1943), and this ratio is a parameter widely used to compare surface energy partitioning (Patil et al., 2011). The Bowen ratios for the BJ and Tanggula sites were both low in Table 4 The monthly and annual averages of the Albedo and Bowen ratio at BJ site and Tanggula site in 2008. BJ

Jan. Feb. Mar. Apr. May. Jun. Jul. Aug. Sep. Oct. Nov. Dec. Avg.

Tanggula

Albedo

Bowen ratio

Albedo

Bowen ratio

0.34 0.40 0.30 0.26 0.23 0.19 0.18 0.18 0.19 0.31 0.73 0.37 0.31

5.66 6.38 18.47 20.98 1.87 0.47 0.37 0.37 0.50 0.93 −0.57 −0.16 1.00

0.37 0.46 0.27 0.28 0.28 0.22 0.20 0.24 0.24 0.58 0.75 0.67 0.38

2.21 0.83 2.80 2.80 0.84 0.54 0.33 0.34 0.33 0.12 −2.00 −1.30 0.53

summer but high in winter and became low or even negative when snow covered the surface. The annual average Bowen ratio was 1.00 and 0.53 for the BJ and Tanggula sites, respectively. The Bowen ratios for the BJ and Tanggula sites were closely related to the seasonal variations in the turbulent fluxes because of the monsoon and to the freezing and thawing processes of the soil, as described in Section 4.2. The monthly average Bowen ratios for the two sites were less than 1.0 from June to September and nearly all exceeded 1.0 in January to April, when the unfrozen soil water content was low and the vegetation coverage was sparse, resulting in the conversion of Rn primarily to Hs. A thick snow cover influenced the Bowen ratio. The monthly average Bowen ratio at the Tanggula site in October was quite low because there was a continuous snow cover after October 14. In November and December, the monthly average Bowen ratio became negative for both sites because the snow cover was thick at both sites. It should be noted that the monthly average Bowen ratios were extremely high for the BJ site in March and April because the calculation of the Bowen ratio has certain limitations under arid surface conditions (Bowen, 1926).

5. Conclusions In this paper, continuous surface energy budgets were obtained at the BJ and Tanggula sites in 2008 with an eddy covariance system and a 10-m meteorological gradient tower data by using the eddy covariance method (with data gap filling by the Bowen ratio and MDV methods). The characteristics of the surface energy budgets for the two sites and their

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

differences were analyzed and compared. The following conclusions were drawn: 1 The energy closure ratios for the BJ and Tanggula sites in 2008 were 0.74 and 0.73, respectively, using instantaneous energy flux data but 0.87 and 0.99 using daily average energy flux data. Therefore, the energy closure status is related to the study's time scale. The monthly CR was relatively poor for the BJ site in November and for the Tanggula site in December because of the thick snow cover. 2 In 2008, the annual averages of Rn, G0, Hs, and LE were 74.73, − 0.50, 34.38, and 34.26 W m− 2, respectively, at the BJ site and 71.59, 0.40, 24.78, and 46.83 W m−2, respectively, at the Tanggula site. The Rn, G0, Hs, and LE exhibited significant seasonal variation: the Rn and G0 were high in summer and autumn but low in winter and spring. The Rn rapidly decreased, or even became negative at both sites when there had been heavy snowfall. From January to March and October to December, the monthly G0 averages at the Tanggula site were lower than those at the BJ site because the surface snow cover was deeper at the Tanggula site than at the BJ site. The Hs was relatively high during spring, decreased in summer, and gradually increased in autumn. The LE showed the opposite trend because the precipitation and water vapor density increase in summer with the monsoon. Simultaneously, the seasonally frozen ground completely thawed, and the active layer of permafrost gradually reached its greatest thawed depth, which increased the moisture content of the soil. In addition, the growth of vegetation flourished, and the soil evapotranspiration increased. All of the above factors caused the LE to increase and predominate at both sites. However, the summer thawing of the active layer consumes considerable energy, which causes the surface warming energy to decline. Therefore, the Hs declined as the land– atmosphere temperature difference declined. Because relatively deep snow covered the surface during the winter in 2008, at both sites, the LE was unusually high, and the Hs was unusually low. 3 Because of the different types of underlying frozen soil, the seasonal variation in turbulent energy flux at the two sites also exhibited different features. The starting time for the significant increase in LE occurred earlier at the Tanggula site than at the BJ site, because the freezing and thawing processes in the permafrost active layer at the Tanggula site significantly affected the Hs and LE distributions, but the same processes at the BJ site did not significantly affect the Hs and LE distributions. The monsoon significantly affected the variation in Hs and LE at both the BJ and Tanggula sites. 4 The Rn, G0, Hs, and LE exhibited significant diurnal variations. The daily maximum Rn at both sites occurred at approximately 14:00 h, which is midday locally. The daily maximum value was earlier for the G0 than for the Hs and LE. The variation of Hs and LE showed seasonally alternating characteristics. 5 The albedo and Bowen ratio for the BJ and Tanggula sites were both low in summer but high in winter. The albedo increased significantly when the surface was snow covered, but the Bowen ratio became low or even negative. The annual average albedo and Bowen ratio were, respectively,

563

0.31 and 1.00 for the BJ site and 0.38 and 0.53 for the Tanggula site. The monthly average albedo was greater for the Tanggula site than for the BJ site. The maximum monthly average albedo occurred in November at both sites. The monthly average Bowen ratios for the two sites were less than 1.0 from June to September and almost all greater than 1.0 from January to April. In November and December, the monthly average Bowen ratio for both sites became negative because of a thick snow cover at both sites.

Acknowledgments This research was supported by the National Key Basic Research Program of China (2013CBA01803, 2010CB951701-2), the National Natural Science Foundation of China (41105004, 40705006, 91337212), and the Innovation Project of the State Key Laboratory of Cryospheric Sciences, CAREERI, CAS (SKLCS 2011-07). References Aubinet, M., Vesala, T., Papale, D. (Eds.), 2012. Eddy Covariance — A Practical Guide to Measurement and Data Analysis. Springer-Verlag, Berlin Heidelberg. Barr, A.G., Kamp, G.V.D., Black, T.A., McCaughey, J.H., Nesic, Z., 2012. Energy balance closure at the BERMS flux towers in relation to the water balance of the White Gull Creek watershed 1999–2009. Agric. For. Meteorol. 153, 3–13. Beringer, J., Chapin III, F.S., Thompson, C.C., McGuire, A.D., 2005. Surface energy exchanges along a tundra–forest transition and feedbacks to climate. Agric. For. Meteorol. 131, 143–161. Bowen, I.S., 1926. The ratio of heat losses by conduction and by evaporation from any water surface. Phys. Rev. 27 (6), 779–787. Cheng, G.D., 2005. Permafrost studies in the Qinghai–Tibet Plateau for road construction. J. Cold Reg. Eng. 19 (1), 19–29. Cong, Z.Y., Kang, S.C., Smirnov, A., Holben, B., 2009. Aerosol optical properties at Nam Co, a remote site in central Tibetan Plateau. Atmos. Res. 92, 42–48. Cui, X.F., Graf, H.F., 2009. Recent land cover changes on the Tibetan Plateau: a review. Clim. Chang. 94, 47–61. Culf, A.D., Foken, T., Gash, J.H.C., 2004. The Energy Balance Closure Problem. Vegetation, Water, Humans and the Climate: A New Perspective on an Interactive System. Springer-Verlag, Berlin Heidelberg. Duan, A.M., Wu, G.X., Liu, Y.M., Ma, Y.M., Zhao, P., 2012. Weather and climate effects of the Tibetan Plateau. Adv. Atmos. Sci. 29, 978–992. Duguay, C.R., Zhang, T.J., Leverington, D.W., Romanovsky, V.E., 2005. Satellite remote sensing of permafrost and seasonally frozen ground. Geophys. Monogr. 163, 91–118. Eugster, W., Rouse, W.R., Pielke, R.A.S.R., Mcfadden, J.P., Baldocchi, D.D., Kittel, T.G.F., Chapin III, F.S., Liston, G.E., Vidale, P.L., Vaganov, E., Chambers, S., 2000. Land–atmosphere energy exchange in Arctic tundra and boreal forest: available data and feedbacks to climate. Global Change Biol. 6 (1), 84–115. Falge, E., Baldocchi, D., Olson, R., Anthoni, P., Aubinet, M., Bernhofer, C., Burba, G., Ceulemans, R., Clement, R., Dolman, H., Granier, A., Gross, P., Grünwald, T., Hollinger, D., Jensen, N.O., Katul, G., Keronen, P., Kowalski, A., Ta, L.C., Law, B.E., Meyers, T., Moncrieff, J., Moors, E., Munger, J.W., Pilegaard, K., Rannik, U., Rebmann, C., Suyker, A., Tenhunen, J., Tu, K., Verma, S., Vesala, T., Wilson, K., Wofsy, S., 2001. Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric. For. Meteorol. 107, 43–69. Finnigan, J.J., Clement, R., Mahli, Y., Leuning, R., Cleugh, H.A., 2003. A reevaluation of long-term flux measurement techniques part I: averaging and coordinate rotation. Bound.-Layer Meteorol. 107, 1–48. Foken, T., 2008a. Micrometeorology. Springer-Verlag, Berlin Heidelberg. Foken, T., 2008b. The energy balance closure problem: an overview. Ecol. Appl. 18 (6), 1351–1367. Foken, T., Wimmer, F., Mauder, M., Thomas, C., Liebethal, C., 2006. Some aspects of the energy balance closure problem. Atmos. Chem. Phys. Discuss 6, 3381–3402. Gu, S., Tang, Y.H., Cui, X.Y., Kato, T., Du, M.Y., Li, Y.N., Zhao, X.Q., 2005. Energy exchange between the atmosphere and a meadow ecosystem on the Qinghai–Tibetan Plateau. Agric. For. Meteorol. 129, 175–185.

564

L. Gu et al. / Atmospheric Research 153 (2015) 553–564

Hendricks-Franssen, H.J., Stöckli, R., Lehner, I., Rotenberg, E., Seneviratne, S.I., 2010. Energy balance closure of eddy-covariance data: a multisite analysis for European FLUXNET stations. Agric. For. Meteorol. 150, 1553–1567. Hirose, N., Koike, T., Ishidaira, H., 2002. Study on spatially averaged evaporation under soil moisture heterogeneity affected by permafrost microtopography. J. Meteorol. Soc. Jpn 80 (2), 191–203. Hsu, H., Liu, X., 2003. Relationship between the Tibetan Plateau heating and East Asian summer monsoon rainfall. Geophys. Res. Lett. 30 (20), 2066. Kaimal, J.C., Finnigan, J.J., 1994. Atmospheric Boundary Layer Flows: Their Structure and Measurement. Oxford University Press, New York. Kalthoff, N., Fiebig-Wittmaack, M., Meißner, C., Kohler, M., Uriarte, M., BischoffGauß, I., Gonzales, E., 2006. The energy balance, evapotranspiration and nocturnal dew deposition of an arid valley in the Andes. J. Arid Environ. 65, 420–443. Leuning, R., Cleugh, H.A., Zegelin, S.J., Hughes, D., 2005. Carbon and water fluxes over a temperate eucalyptus forest and a tropical wet/dry savanna in Australia: measurements and comparison with MODIS remote sensing estimates. Agric. For. Meteorol. 129, 151–173. Leuning, R., Gorsel, E.V., Massman, W.J., Isaac, P.R., 2012. Reflections on the surface energy imbalance problem. Agric. For. Meteorol. 156, 65–74. Li, C.L., Kang, S.C., Zhang, Q.G., Kaspari, S., 2007. Major ionic composition of precipitation in the Nam Co region, Central Tibetan Plateau. Atmos. Res. 85, 351–360. Liu, H.P., Peters, G., Foken, T., 2001. New equations for sonic temperature variance and buoyancy heat flux with an omnidirectional sonic anemometer. Bound.-Layer Meteorol. 100, 459–468. Ma, Y.M., Tsukamoto, O., 2002. Combining Satellite Remote Sensing With Field Observations for Land Surface Heat Fluxes Over Inhomogeneous Landscape. China Meteorological Press, Beijing. Ma, Y.M., Su, Z.B., Koike, T., Yao, T.D., Ishikawa, H., Ueno, K., Menenti, M., 2003. On measuring and remote sensing surface energy partitioning over the Tibetan Plateau—from GAME/Tibet to CAMP/Tibet. Phys. Chem. Earth 28, 63–74. Ma, Y.M., Kang, S.C., Zhu, L.P., Xu, B.Q., Tian, L.D., Yao, T.D., 2008. Tibetan observation and research platform: atmosphere–land interaction over a heterogeneous landscape. Bull. Am. Meteorol. Soc. 89, 1487–1492. Ma, Y.M., Wang, Y.J., Wu, R.S., Hu, Z.Y., Yang, K., Li, M.S., Ma, W.Q., Zhong, L., Sun, F.L., Chen, X.L., Zhu, Z.K., Wang, S.Z., Ishikawa, H., 2009. Recent advances on the study of atmosphere–land interaction observations on the Tibetan Plateau. Hydrol. Earth Syst. Sci. 13, 1103–1111. Ma, Y.M., Zhong, L., Wang, Y.J., Su, Z.B., 2012. Using NOAA/AVHRR data to determine regional net radiation and soil heat fluxes over the heterogeneous landscape of the Tibetan Plateau. Int. J. Remote Sens. 33, 4784–4795. Ma, W.Q., Ma, Y.M., Ishikawa, H., 2014. Evaluation of the SEBS for upscaling the evapotranspiration based on in-situ observations over the Tibetan Plateau. Atmos. Res. 138, 91–97. Massman, W.J., 2000. A simple method for estimating frequency response corrections for eddy covariance systems. Agric. For. Meteorol. 104, 185–198. Mauder, M., Cuntz, M., Drüe, C., Graf, A., Rebmann, C., Schmid, H.P., Schmidt, M., Steinbrecher, R., 2013. A strategy for quality and uncertainty assessment of long-term eddy-covariance measurements. Agric. For. Meteorol. 169, 122–135. Oncley, S.P., Foken, T., Vogt, R., Kohsiek, W., DeBruin, H.A.R., Bernhofer, C., Christen, A., Gorsel, E.V., Grantz, D., Feigenwinter, C., Lehner, I., Liebethal, C., Liu, H.P., Mauder, M., Pitacco, A., Ribeiro, L., Weidinger, T., 2007. The energy balance experiment EBEX-2000, part I: overview and energy balance. Bound.-Layer Meteorol. 123, 1–28. Patil, M.N., Waghmare, R.T., Halder, S., Dharmaraj, T., 2011. Performance of Noah land surface model over the tropical semi-arid conditions in western India. Atmos. Res. 99, 85–96. Pavlov, A.V., 1994. Current change of climate and permafrost in the Arctic and sub-Arctic of Russia. Permafr. Periglac. Process. 5, 101–110. Ramos, M., Vieira, G., Gruber, S., Blanco, J.J., Hauck, C., Hidalgo, M.A., Tomé, D., Neves, M., Trindade, A., 2007. Permafrost and active layer monitoring in the maritime Antarctic: preliminary results from CALM sites on Livingston and Deception Islands. USGS OF-2007-1047, Short Research Paper 070. U.S. Geological Survey and The National Academies. Roger, B., Thian, Y.G., 2011. The Global Cryosphere: Past, Present and Future. Cambridge University Press, Cambridge. Sato, T., Kimura, F., 2007. How does the Tibetan Plateau affect the transition of Indian monsoon rainfall? Mon. Weather Rev. 135, 2006–2015. Shi, Y.F., Yu, G., Liu, X.D., Li, B.Y., Yao, T.D., 2001. Reconstruction of the 30–40 ka BP enhanced India monsoon climate based on geological record from the Tibetan Plateau. Palaeogeogr. Palaeoclimatol. Palaeoecol. 169, 69–83.

Stoy, P.C., Mauder, M., Foken, T., Marcolla, B., Boegh, E., Ibrom, A., Arain, M.A., Arneth, A., Aurela, M., Bernhofer, C., Cescatti, A., Dellwik, E., Duce, P., Gianelle, D., Gorsel, E.V., Kiely, G., Knohl, A., Margolis, H., McCaughey, H., Merbold, L., Montagnani, L., Papale, D., Reichstein, M., Saunders, M., Serrano-Ortiz, P., Sottocornola, M., Spano, D., Vaccari, F., Varlagin, A., 2013. A data-driven analysis of energy balance closure across FLUXNET research sites: the role of landscape scale heterogeneity. Agric. For. Meteorol. 171–172, 137–152. Strugnell, N.C., Lucht, W., 2001. An algorithm to infer continental-scale albedo from AVHRR data, land cover class, and field observations of typical BRDFs. J. Clim. 14, 1360–1376. Sverdrup, H.U., 1943. On the ratio between heat conduction from the sea surface and the heat used for evaporation. Ann. N. Y. Acad. Sci. 68, 81–88. Tanaka, K., Ishikawa, H., Hayashi, T., Tamagawa, I., Ma, Y.M., 2001. Surface energy budget at Amdo on the Tibetan Plateau using GAME/Tibet IOP98 data. J. Meteorol. Soc. Jpn 79 (1B), 505–517. Tanaka, K., Tamagawa, I., Ishikawa, H., Ma, Y.M., Hu, Z.Y., 2003. Surface energy budget and closure of the eastern Tibetan Plateau during the GAME-Tibet IOP 1998. J. Hydrol. 283, 169–183. Webb, E.K., Pearman, G.I., Leuning, R., 1980. Correction of flux measurements for density effects due to heat and water vapor transfer. Q. J. R. Meteorol. Soc. 106, 85–100. Wilson, K., Goldstein, A., Falge, E., Aubinet, M., Baldocchi, D., Berbigier, P., Bernhofer, C., Ceulemans, R., Dolman, H., Field, C., Grelle, A., Ibrom, A., Law, B.E., Kowalski, A., Meyers, T., Moncrieff, J., Monson, R., Oechel, W., Tenhunen, J., Valentini, R., Verma, S., 2002. Energy balance closure at FLUXNET sites. Agric. For. Meteorol. 113, 223–243. Woo, M.K., 2012. Permafrost Hydrology. Springer-Verlag, Berlin Heidelberg. Xu, X.D., Lu, C.G., Shi, X.H., Gao, S.T., 2008. World water tower: an atmospheric perspective. Geophys. Res. Lett. 30 (20) (L20815). Xue, B.L., Wang, L., Li, X.P., Yang, K., Chen, D.L., Sun, L.T., 2013. Evaluation of evapotranspiration estimates for two river basins on the Tibetan Plateau by a water balance method. J. Hydrol. 492, 290–297. Yanai, M., Li, C.F., Song, Z.S., 1992. Seasonal heating of the Tibetan Plateau and its effects on the evolution of the Asian summer monsoon. J. Meteorol. Soc. Jpn 70, 319–351. Yang, M.X., Yao, T.D., Gou, X.H., Koike, T., He, Y.Q., 2003. The soil moisture distribution, thawing–freezing processes and their effects on seasonal transition on the Qinghai–Xizang (Tibetan) Plateau. J. Asian Earth Sci. 21, 457–465. Yang, K., Ye, B.S., Zhou, D.G., Wu, B.Y., Foken, T., Qin, J., Zhou, Z.Y., 2011. Response of hydrological cycle to recent climate changes in the Tibetan Plateau. Clim. Chang. 109, 517–534. Yang, K., Wu, H., Qin, J., Lin, C.G., Tang, W.J., Chen, Y.Y., 2014. Recent climate changes over the Tibetan Plateau and their impacts on energy and water cycle: a review. Global Planet. Chang. 112, 79–91. Yao, J.M., Zhao, L., Ding, Y.J., Gu, L.L., Jiao, K.Q., Qiao, Y.P., Wang, Y.X., 2008. The surface energy budget and evapotranspiration in the Tanggula region on the Tibetan Plateau. Cold Reg. Sci. Technol. 52, 326–340. Yao, J.M., Zhao, L., Gu, L.L., Qiao, Y.P., Jiao, K.Q., 2011. The surface energy budget in the permafrost region of the Tibetan Plateau. Atmos. Res. 102 (4), 394–407. Ye, D.Z., 1981. Some characteristics of the summer circulation over the Qinghai– Xizang (Tibet) Plateau and its neighborhood. Bull. Am. Meteorol. Soc. 62, 14–19. Ye, D.Z., Gao, Y.X., 1979. The Meteorology of the Qinghai–Xizang (Tibet) Plateau. Science Press, Beijing (in Chinese). Ye, D.Z., Wu, G.X., 1998. The role of the heat source of the Tibetan Plateau in the general circulation. Meteorol. Atmos. Phys. 67, 181–198. Yu, W.S., Yao, T.D., Tian, L.D., Ma, Y.M., Ichiyanagi, K., Wang, Y., Sun, W.Z., 2008. Relationships between δ18O in precipitation and air temperature and moisture origin on a south–north transect of the Tibetan Plateau. Atmos. Res. 87, 158–169. Zhao, L., Cheng, G.D., Li, S.X., Zhao, X.M., Wang, S.L., 2000. Thawing and freezing processes of active layer in Wudaoliang region of Tibetan Plateau. Chin. Sci. Bull. 45 (23), 2181–2187. Zhao, L., Ping, C.L., Yang, D.Q., Cheng, G.D., Ding, Y.J., Liu, S.Y., 2004. Changes of climate and seasonally frozen ground over the past 30 years in Qinghai– Xizang (Tibetan) Plateau, China. Global Planet. Chang. 43, 19–31. Zheng, W., Yao, T.D., Joswiak, D.R., Xu, B.Q., Wang, N.L., Zhao, H.B., 2010. Major ions composition records from a shallow ice core on Mt. Tanggula in the central Qinghai–Tibetan Plateau. Atmos. Res. 97, 70–79. Zhou, Y.W., Guo, D.X., Qiu, G.Q., Cheng, G.D., Li, S.D., 2000. Geocryology in China. Science Press, Beijing (in Chinese).