-
Journal of Hydrology: Regional Studies 5 (2016) 20–32
Contents lists available at ScienceDirect
Journal of Hydrology: RegionalStudies
jo ur nal homep age: www.elsev ier .com/ locate /e j rh
Discriminating types of precipitation in Qilian
Mountains,Tibetan Plateau
Junfeng Liua,b,∗, Rensheng Chena,b
a Qilian Alpine Ecology and Hydrology Research Station, Cold and
Arid Regions Environmental and Engineering Research
Institute,Chinese Academy of Sciences, Lanzhou 730000, Chinab Heihe
Key Laboratory of Ecohydrology and Integrated River Basin
Management, Chinese Academy of Sciences, Lanzhou 730000, China
a r t i c l e i n f o
Article history:Received 4 May 2015Received in revised form 1
November 2015Accepted 15 November 2015Available online 28 November
2015
Keywords:Precipitation typeThreshold air temperatureHulugou
river basinQilian MountainousTibetan Plateau
a b s t r a c t
Study region: Hulugou River Basin (HRB) of Qilian Mountains,
eastern edge of TibetanPlateau.Study focus: Traditional manual
observations only record point-scale precipitation ratherthan
regional-scale precipitation. Automatic weather stations just
record precipitationamount without discriminating by type of
precipitation. This study observed precipita-tion types all over
the HRB and analyzed air temperature and humidity conditions at
dailyand half-hour resolution.New hydrological insights for the
region: Combined observations of air temperature and pre-cipitation
type indicate that, at daily resolution the threshold air
temperature betweenrain and snow is 0 ◦C and the air temperature at
rain/snow boundary is from 0 ◦C to 7.6 ◦C,which means the rain and
mixed precipitation threshold air temperature can shift morethan
7.0 ◦C in the HRB. At half-hour resolution, air temperature is
above 0 ◦C during rainfall,under 0 ◦C for snowfall, and above 0 ◦C
at the rain/snow precipitation boundary, and couldeither be above 0
◦C or fluctuate around 0 ◦C for mixed precipitation. Corresponding
rela-tive humidity observations indicate that rainfall and mixed
precipitation events correspondwith high humidity conditions in
warm season of the HRB. Snowfall events correspond withlow humidity
conditions in the HRB.© 2015 The Authors. Published by Elsevier
B.V. This is an open access article under the CC
BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Observation of precise precipitation amount and type of
precipitation are critical for reliable stream flow
forecasts(Fassnacht et al., 2006). Precipitation amount has been
widely observed, but precipitation type is rarely observed by
auto-matic weather stations. Type of precipitation has great impact
on the land surface energy balance (Loth et al., 1993; Dinget al.,
2014), as snowfall can increase the surface albedo, and rainfall
can decrease the surface albedo (Box et al., 2012). Land
surface hydrological processes are also different for different
precipitation types, i.e., rainfall usually converts into
under-ground or surface water stream systems, while snowfall may
accumulate during the cold season and melt in warm
season.Observation of precipitation type is important, because some
forms of cold season precipitation can pose a threat to humansafety
or disrupt travel and commerce (Reeves et al., 2014). It has been
widely recognized that gauge measured precipitation
∗ Corresponding author at: Qilian Alpine Ecology and Hydrology
Research Station, Cold and Arid Regions Environmental and
Engineering ResearchInstitute, Chinese Academy of Sciences, Lanzhou
730000, China
E-mail address: [email protected] (J. Liu).
http://dx.doi.org/10.1016/j.ejrh.2015.11.0132214-5818/© 2015 The
Authors. Published by Elsevier B.V. This is an open access article
under the CC BY-NC-ND
license(http://creativecommons.org/licenses/by-nc-nd/4.0/).
dx.doi.org/10.1016/j.ejrh.2015.11.013http://www.sciencedirect.com/science/journal/22145818http://www.elsevier.com/locate/ejrhhttp://crossmark.crossref.org/dialog/?doi=10.1016/j.ejrh.2015.11.013&domain=pdfhttp://creativecommons.org/licenses/by-nc-nd/4.0/mailto:[email protected]/10.1016/j.ejrh.2015.11.013http://creativecommons.org/licenses/by-nc-nd/4.0/
-
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 21
Table 1Precipitation measurements intercomparison
experiment.
Gauge Abbreviation Size Start date End date Measure time
China standard precipitation gauge CSPG � =20 cm, h = 70 cm
June, 2009 April, 2014 20:00 and 8:00, local timeCSPG with Alter
shelter Alter � = 20 cm, h = 70 cm June, 2009 April, 2014 20:00 and
8:00, local timePit gauge with a CSPG Pit � = 20 cm, h = 0 cm
September, 2010 April, 2014 20:00 and 8:00, local timePit with a
larger diameter bucket Pit500 � = 25 cm, h = 0 cm September, 2010
April, 2014 20:00 and 8:00, local time
htd
aMIatcrvo
vtitpbihmb
2
9tsantaa
HranwOtC1aH
A
Double-Fence International Reference(Tretyakov wind shield +
CSPG)
DFIR � = 20 cm, h = 3.0 m September, 2012 April, 2014 20:00 and
8:00, local time
as systematic errors mainly caused by wind-induced undercatch
(Sugiura et al., 2003). When the wind speed is very high,he catch
ratio of precipitation gauges depends on precipitation type (Yang
et al., 1995). In this case, the precipitation typeata deficit will
greatly affect measured precipitation calibration and
precision.
However, most of the time precipitation type data are not
available because these depend on manual observation andutomatic
precipitation type discrimination sensors are not precise enough to
be used widely. For example, the Worldeteorological Organization
arranged an intercomparison of weather sensor abilities (Present
Weather Sensors/Systems
ntercomparison; PREWIC, 1993–1995) in the field of precipitation
detection and precipitation type discrimination (Leroynd Bellevaux,
1998). Sensors like the Vaisala FD12P performed the best of all
optical sensors in PREWIC, but it is known thathe discrimination of
precipitation type is not precise (Haji, 2007). Manual observations
of precipitation type are limited atertain altitudes. A study by
Thériault et al. (2010) indicated that small variations in
temperature profiles and precipitationate can have major impacts on
the types of precipitation formed at the surface. It was also found
that precipitation typearies widely over space and time in
mountainous regions (Harder and Pomeroy, 2013), where the rain/snow
boundaryften formed at certain altitudes.
Little knowledge is available of how elevation impact the
precipitation types, and no systemic precipitation type
obser-ations have been undertaken at watershed scale in mountainous
regions. To address this shortfall, a systemic precipitationype
observation experiment was carried out in the Hulugou River Basin
(HRB) of the Qilian Mountainous. The observationsnclude manual
observations since 2009 as a reference at the outlet of the HRB.
Two time-lapse cameras have been usedo photograph precipitation
types in the HRB at two different locations since 2012. Through
photographic interpretation,recipitation types which include
rainfall, snowfall and mixed precipitation are acquired at basin
scale. If the rain/snowoundary is visualized in a precipitation
event, the elevation of the rain/snow boundary is obtained by
georeferencing the
mage. On the basis of automatic land surface meteorological
observations at different altitudes, the air temperature andumidity
were determined for each precipitation event at daily and half-hour
resolution. Specifically, in this study, theeteorological
conditions at the rain/snow boundary were determined during each
mixed precipitation event, which can
e used as the threshold air temperature to differentiate rain
and mixed precipitation.
. Study area and data
The HRB is located in the Qilian Mountains along the
northeastern margin of the Tibetan Plateau
(38◦15′54.9′′N,9◦52′53.5′′E; Fig. 1). The basin forms the
headwaters of the Heihe River and has a catchment area of 23.4 km2.
Eleva-ions in the basin range from 2980 m to 4800 m above sea
level. The landscape zones include glaciers at mountain
top,ubnival, marsh meadows, alpine shrub, and mountain grassland.
The annual precipitation ranges from 376 to 650 mm,nd the annual
mean temperature varies from approximately 3.1 ◦C at 3000 m to −4.0
◦C at 4200 m. The river basin facesorth with inclines from 5◦ to
85◦. The steep north facing slope experiences less solar radiation
and lower ground surfaceemperatures, combined with the large
elevation ranges, these conditions cause mixed precipitation events
to frequentlyppear at high altitudes during the warm season.
Consequently, the rain/snow boundary is often visualized on north
facingspects of the HRB.
Field precipitation measurement intercomparison experiments
(Fig. 1, Table 1) were conducted near the outlet of theRB (2980 m).
Manual observations (rain, snow and mixed) are three time per day.
In September 2008, four ENVISs (Envi-
onmental Information Systems) were installed at 2980 m, 3382 m,
3710 m, and 4166 m in the HRB. Measurements includeir temperature,
humidity, wind speed and direction, radiation, amount of
precipitation (weighing gauge), snow depth,umber of hours of sun,
ground surface temperature, soil temperature at eight depths, and
three ground heat fluxes. Theeighting gauge was used to determine
the total rain using a weighting sensor TRwS500 with a resolution
of 0.001 mm.ther sensors installed in each of the ENVISs include
the temperature and relative humidity sensor of Campbell
CS215-L,
he wind speed and direction sensor of Campbell Wind sonic
Two-Dimensional Sonic Anemometer, the radiation sensor ofampbell
four-component radiometer, the Campbell snow depth sensor of SR50
Sonic Ranging Sensor with a precision of
cm, the Campbell soil water content sensor of Enviro SMART Soil
Water Content Profile Probes, the Campbell soil temper-ture sensor
of 107 L Sensor, the CSD3sun time hour sensor, and the soil heat
flux sensor of HFP01SC-L Self-Calibrating Soil
eat Flux Plate sensor.
In June 2012, two automatic weather stations (AWSs) were
installed on the western tributary of the HRB (3839 m).nother AWS
was installed at a higher site (4496 m) on the Shiyi glacier. Each
AWS measures air temperature, relative
-
22 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
Fig. 1. Overview of the Hulugou River Basin. Shown are the
locations of observation points of the manual precipitation, two
automatic weather stations,four environmental information systems,
the Shiyi glacier, the EOS 7D camera at 3140 m and the EOS 600D
camera at 4550 m.
humidity, wind speed and precipitation once every 30 min. The
sensors in these two AWSs for these measurements weresimilar to
those in the ENVISs.
Since September 2011, manual shoot has been applied in the HRB
at 3140 m every morning around 9:00 Beijing time,approximately 1 km
from the catchment outlet. In April 2012, a Canon EOS 7D was
equipped with automatic time-lapseaccessories and a protective
device to replace manual shooting. This camera was installed at
manual shooting site. Before
July 6 2012, pictures were taken automatically at 9:51 and 15:51
Beijing time, and after June 1, the pictures were taken onthe hour.
The camera’s field of view covers the whole basin’s altitudinal
gradient. This camera records the precipitation typeand snow cover
distribution for the whole HRB.
-
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 23
Convert to orthophotos
Snow
Manual observa�on of p recipita�on type at the HRB o utlet at
2980 m
Snow for the HRB
Rain Formed n ew sn ow/rain bounda ry?No
Camera Hourly Photography
Extract al�tude of rain/snow boundary
Rain for the HRBYes
Rain under rain/snow boundary, mixed precipita�on above
Mixed precipita�on
Mixed precipita�on for the HRB
wtpagbc
3
0ottfmb
tiotpab
ot
e
o
p
dp
3tasm
Fig. 2. Flowchart describing the processes used to discriminate
precipitation types in the Hulugou River Basin (HRB).
In July 2012, a Canon EOS 600D time-lapse camera was installed
at 4550 m. The camera photographs the Shiyi glacier,hich covers 0.3
km2 and ranges from 4313 m at its terminus to 4800 m above sea
level. In this study, this camera was used
o record high altitude precipitation types. The glacier surface
temperatures are equal to or less than 0 ◦C, which
meansrecipitation type is hardly affected by the glacier surface
thermal conditions. In the warm season, clouds or fog
persistentlyffect the view of the EOS 7D camera. Compared to the
EOS7D camera, the EOS 600D camera has a closer view of the
Shiyilacier and can sometimes avoid the clouds for fog. Sometimes,
if the solid parts are low, the snow/rain boundary can hardlye
observed by the EOS7D camera, but the EOS600D can detect it because
of its close view. In this case, the EOS 600D cameraan observe more
mixed precipitation than the EOS 7D camera.
. Methods
In the HRB cold season (October–April) observed air temperatures
at 2980 m, in the basin outlet, are usually under◦C. Precipitation
intensity and duration are low most of the time, and the
precipitation amount is usually less than 30%f the annual total.
The manual observation at 2980 m indicates that snowfall dominates
the cold season precipitationhroughout the HRB. Precipitation is
mainly concentrated in the warm season. Precipitation types can
change from raino mixed precipitation or snowfall with increasing
altitude. Even at the same altitude, precipitation may be preceded
orollowed by mixed precipitation. In this case, precipitation type
discrimination is a very difficult task. In the HRB, most of
the
ixed precipitation or snowfall events happened at night, but
precipitation type could only be monitored during daytimey
observers or cameras after it happened.
Fig. 2 shows a flowchart of precipitation type discrimination
methods used in this paper. The discrimination of precipi-ation is
based on manual observations at the basin outlet at 2980 m and two
cameras at 3140 m and 4550 m. In this study,f the manual
observation at the basin outlet is snow, or mixed precipitation,
the entire HRB is considered to have snowfallr mixed precipitation.
If the basin outlet manual observation is rainfall, the two cameras
were used to identify precipita-ion type at higher altitudes, where
if either the EOS7D camera or EOS 600D camera captures the
rain/snow boundary, therecipitation under the rain/snow boundary is
considered to be rainfall and above the boundary is mixed
precipitation. Theltitude of the rain/snow boundary was calculated
based on georeferencing these pictures. The georeferencing method
haseen presented by Liu et al. (2012, 2014).
In order to avoid subjective discrimination of precipitation
type at high altitudes on the Shiyi glacier, we also used
thebserved snow depth data, half-hour precipitation data, half-hour
air temperature, and pictures taken by the two cameraso help in the
precipitation type discrimination process.
The three TRwS500 weighing gauges at 3340 m, 3710 m and 4160 m
are a cumulative record of both rainfall and snowfallvents. Phase
was identified if the TRws500 recorded precipitation.
Snow depth was measured by the SR50 sensor. An increase in snow
depth was used to identify snowfall, while no changer a decrease in
snow depth was used to identify rainfall.
Ground surface temperature was used to identify snowfall and
rainfall in the warm season. If the ground surface tem-erature was
greater than 0 ◦C, then all precipitation was presumed to be
rainfall.
If the two cameras captured snowfall, or the snow depth sensor
recorded an increase in snow depth, the density of snowepth was
calculated based on observed precipitation data and snow depth
data. If snow density was over 0.2 g/cm3, therecipitation was
identified as mixed, while a density under 0.2 g/cm3 was identified
as snowfall.
Fig. 3 presents a mixed precipitation event starting on May 23,
2012 with associated meteorological data at 3340 m,710 m and 4160
m. At 3340 m, the snow depth was not captured by the SR50 snow
depth sensors. At 3710 m and 4160 m,
he air temperature and ground surface temperature decreased with
altitude, and snow depth and duration increased withltitude. During
this precipitation event, the EOS 7D camera recorded the snow/rain
boundary at 3185 m, which means thenow depth sensor could not
detect the thin snow depth. At 3710 m and 4160 m, the density was
over 0.2 g/cm3, whicheans mixed precipitation happened.
-
24 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
0
0.5
1
1.5
2
2.5
3
-3
-1
1
3
5
7
9
11
Dat
e
5/23
1:0
0
5/23
2:3
0
5/23
4:0
0
5/23
5:3
0
5/23
7:0
0
5/23
8:3
0
5/23
10:
00
5/23
11:
30
5/23
13:
00
5/23
14:
30
5/23
16:
00
5/23
17:
30
5/23
19:
00
5/23
20:
30
5/23
22:
00
Pre
cipi
tati
on/m
m a
nd s
now
dep
th/c
m
/erutarepme
To C
ba
c
Prec ipitation (mm )Ground surface temperatureAir
temperature
0
0.5
1
1.5
2
2.5
3
3.5
-4
-2
0
2
4
6
8
Dat
e
5/23
1:0
0
5/23
2:3
0
5/23
4:0
0
5/23
5:3
0
5/23
7:0
0
5/23
8:3
0
5/23
10:
00
5/23
11:
30
5/23
13:
00
5/23
14:
30
5/23
16:
00
5/23
17:
30
5/23
19:
00
5/23
20:
30
5/23
22:
00
Pre
cipi
tati
on/m
m a
nd s
now
dep
th/c
m
/erutarepme
To C
Prec ipitation (mm )Grou nd surface temperatureAir
temperature
Snow depth (cm)
0
1
2
3
4
5
6
7
8
-6
-4
-2
0
2
4
6
8
Dat
e
5/23
1:0
0
5/23
2:3
0
5/23
4:0
0
5/23
5:3
0
5/23
7:0
0
5/23
8:3
0
5/23
10:
00
5/23
11:
30
5/23
13:
00
5/23
14:
30
5/23
16:
00
5/23
17:
30
5/23
19:
00
5/23
20:
30
5/23
22:
00
Pre
cipi
tati
on/m
m a
nd s
now
dep
th/c
m
/er utarepme
To C
Prec ipitationGround surface temperatureAir temperature
Snow depth
Fig. 3. Observed snow depth and associated meteorological data
at three different altitudes: (a) 3340 m, (b) 3710 m and (c) 4160
m.
4. Results
4.1. Precipitation type observation results at three
locations
On the basis of observations at three different locations, great
spatial and temporal differences in precipitation typeswere
observed in the HRB (Fig. 4). Manual observations at the HRB outlet
showed that rain accounted for 69% and 83% ofthe precipitation
events in 2012 and 2013, respectively; mixed precipitation events
accounted for 2.4% and 5%, snow eventsaccounted for 29% and 12% in
2012 and 2013. Rainfall was mainly concentrated from April to
September, and snowfall mainlyoccurred from October to March. At
4550 m, observed rain, mixed and snow precipitation events occur
respectively 33%,38%, and 29% in 2013. Compared with the manual
observation results at the basin outlet, mixed precipitation events
arevery common phenomena and rain occurs less often in the warm
season at high altitudes above 4000 m.
Compared with the manual observations at 2980 m or the EOS 600D
camera results at 4550 m, the EOS 7D cameraprovides a basin-scale
perspective for observing precipitation type. For example, on 18
July 2013, the manual observation at2980 m was rain, and a heavy
snowfall covered the whole Shiyi glacier and surrounding area at
4550 m (Fig. 5). The EOS 7Dcamera caught the rain/snow boundary at
3600 m (Fig. 5). Under this circumstance, the basin-scale
precipitation is classifiedas mixed rather than rain or snow. Only
when the whole HRB is experiencing rain or snow, the precipitation
is classifiedaccordingly, otherwise it is classified as mixed
precipitation. Fig. 6 shows the observed and monitored
precipitation type atthree different locations in 2012 and 2013,
based on this discrimination principle. Mixed precipitation
appeared from Aprilto October in 2012, and accounts for
approximately 50% of the precipitation events. Basin-scale rainfall
events occurredonly 12 times in July and August of 2012, and snow
events occurred 32 times in whole year of 2012. In 2013, rain
eventshappened 36 times from June to August. Mixed precipitation
happened 58 times from April to October, and snow occurred12 times
in whole year of 2013.
After georeferencing the basin-scale mixed and snow
precipitation events, we get the altitude of the rain/snow
boundary(Fig. 7), where rain changed to mixed precipitation. In
Fig. 7a it can be seen that from April to July 2012, the
rain/snowboundary altitude increased to greater than 4000 m. The
elevation of the rain/snow boundary then decreased to 3000 m
inOctober. The change in elevation is closely related to the
seasonal change in air temperatures. Fig. 7b shows that the
warmseason rain/snow boundary altitude variation in 2013 is similar
to 2012, except in July and August, during which the EOS 7Dcamera
only caught the rain/snow boundary once in 2013. In July and August
of 2013, more precipitation events occurred
than in the same months of 2012. The EOS 600D camera caught 14
and 13 mixed precipitation events from July and Augustin 2012 and
2013, respectively, but fewer rain/snow boundaries were caught by
the EOS 7D camera in 2013, which meansthe rain/snow boundary had
moved to even higher altitudes.
-
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 25
0
2
4
6
8
10
12
14
16
stnevenoi tatipicerP
mon th
a snow: 24mixed: 2
rain: 57
0
5
10
15
20
25
stnevenoitatipic erp
mon th
b sno w: 12mixed: 5rain: 84
02468
1012141618
stnevenoitatipicerP
mon th
c snow: 32mixed: 44rain: 14
0
5
10
15
20
25
30
stnevenoitatipicerp
month
dsnow: 12mixed: 58rain: 36
02468
1012141618
stne venoitati pi cer P
month
e snow: 38mixed: 38rain: 16
0
5
10
15
20
25
30
stnevenoit at ipi cerp
mon th
f sno w: 32mixed: 42rain: 37
Fig. 4. Observed number of events for rain, mixed precipitation,
or snow at three different locations in 2012 and 2013. (a) and (b)
show manual observationsat the HRB outlet at 2980 m in 2012 and
2013, respectively. (c) and (d) show EOS 7D camera observation
results for the whole basin in 2012 and 2013,respectively. (e) and
(f) show EOS 600D camera observation results for precipitation type
on the Shiyi glacier surface at 4550 m in 2012 and 2013,
respectively.
dopcmtHo
2(aaEap
From basin-scale comparative studies of precipitation type over
two successive years, we find that mixed precipitationominated the
HRB precipitation events from April to October. From October to
March, snow dominated. Rain mainlyccurred in July and August.
Precipitation type proportions were very different in 2012 and
2013. In 2012, rain, mixedrecipitation and snow accounted for 16%,
49% and 35% of precipitation events, respectively; but in 2013, the
proportionshanged to 34%, 55% and 11%, respectively. Air
temperature is generally the determining factor for the
partitioning of rain,ixed and snow (Braun, 1985). More rain and
less snow in 2013 indicated higher air temperatures. Comparative
studies at
wo different altitudes in 2012 and 2013 indicated that monthly
air temperatures in 2013 were higher than in 2012 in theRB (Fig.
8). In July and August 2013, the air temperatures were 0.4 ◦C and
1.4 ◦C higher than in the corresponding monthsf 2012. Consequently,
the rain proportion was much higher in 2013 than in 2012.
Another difference from 2012 to 2013 is that basin-scale snow
events lasted until late May in 2013, much later than in012 (when
they stopped in April), and the mean rain/snow boundary altitude
was lower in May of 2013 than in May of 2012Fig. 7). Detailed
studies of these precipitation events indicated that in May of
2013, three heavy mixed precipitation eventsnd one small mixed
precipitation event happened across the entire HRB, and the
corresponding air temperatures fluctuatedround 0 ◦C. In May of 2012
only two small mixed precipitation events were noted in manual
observations at the basin outlet.ven though the monthly air
temperature was not different in May of 2012 than May of 2013, the
higher rain/snow boundarynd fewer basin-scale mixed precipitation
events in May of 2012 were related to the higher air temperatures
during therecipitation period.
-
26 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
Fig. 5. Precipitation type observations at 3140 m and 4550 m on
18 July 2013. (a) and (b) were photographed by EOS 7D camera over
basin scale. (a) wastaken before mixed precipitation occurred, and
(b) was taken after mixed precipitation occurred. (c) and (d) were
photographed by EOS 600D camera onthe Shiyi glacier. (c) was taken
before snowfall occurred, and (d) was after snowfall occurred.
4.2. Daily mean threshold air temperatures in the HRB
Static or dynamic threshold air temperature has been widely
applied to differentiate rain, mixed precipitation, or snow.We used
observed air temperatures at different altitudes and precipitation
type results to obtain the threshold air tempera-tures. Fig. 9
shows precipitation types and corresponding daily air temperatures.
Manual observations at 2980 m indicatedthat rain appeared when the
air temperature was above 3.6 ◦C, snow was not observed when the
air temperature exceeded−2.7 ◦C, and mixed precipitation appeared
when the air temperature was between 1.9 ◦C and 6.1 ◦C (Fig. 9a).
Results from
the EOS 600D camera at 4550 m indicated that rain appeared when
the air temperature was above 1.7 ◦C, snow appearedwhen the air
temperature was under −0.5 ◦C, and mixed precipitation appeared
when the air temperature was between 0 ◦Cand 7.6 ◦C (Fig. 9b).
0
20
40
60
80
100
120
2012 201 2 201 2 2013 201 3 201 3
Basin outlet
EOS 7D Shiyi gla cier
Basin outlet
EOS 7D Shiyi glacier
stnevenoitatipicerp
yea r and locati ons
snow: 16 2 mixed: 17 7 Rain: 24 4
Fig. 6. Observed precipitation type at three different locations
in 2012 and 2013.
-
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 27
-30
-25
-20
-15
-10
-5
0
5
10
15
2900
3100
3300
3500
3700
3900
4100
4300
4500
01/01 /12 03/31/12 06 /29 /12 09/27/12 12/26/12
air t
empe
ratu
re (
o C)
)m (
edut itla
Date
a:2012sno w li ne alti tude (m)air temperature at 4550 m ( )
-30
-25
-20
-15
-10
-5
0
5
10
15
2900
3100
3300
3500
3700
3900
4100
4300
4500
01/01 /13 04/01/13 06 /30 /13 09/28/13 12/27/13
air t
empe
ratu
re (
o C)
eduti tla(m
)
Date
b: 20 13snow line altitude (m)air temperature at 4550 m ( )
ba
Fig. 7. Rain/snow boundary altitudes in (a) 2012 and (b) 2013,
caught with the EOS 7D camera; and daily air temperatures at 4550 m
(rain/snow boundaryaltitude equal to 2980 m means basin scale snow
events or mixed precipitation; for rain/snow boundary altitudes
greater than 2980 m, below the boundaryis rain and above is mixed
precipitation).
Fig. 8. Observed monthly air temperatures at (a) 2980 m and (b)
4550 m in 2012 and 2013 in the Hulugou River Basin.
Fig. 9. Rain, mixed precipitation and snow precipitation
threshold mean daily air temperatures (a) air temperature at 2980
m, (b) air temperature at4550 m, (c) air temperature at rain/snow
boundary for mixed rain, air temperature at 4765 m for rain, and
air temperature at 2980 m for snow).
-
28 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
Fig. 10. Half-hour air temperatures during rain, mixed
precipitation and snow events, and air temperatures at the
rain/snow boundary during precipitationevents (a) half-hour air
temperatures at 4550 m for rain events, (b) half-hour air
temperatures during precipitation events at the rain/snow boundary,
(c)half-hour air temperatures during mixed precipitation events,
(d) half-hour air temperatures at 2980 m for snow events).
To calculate the threshold air temperature at the rain/snow
boundary, first, the images were georeferenced into a
digitalorthophoto with the same resolution as the available digital
elevation model data (DEM). Then, the altitude at the
rain/snowboundary is extracted based on the orthophoto and DEM
data. Finally, two observed air temperatures that are close tothe
rain/snow boundary were selected, and half-hour air temperatures
were interpolated to the altitude of the rain/snowboundary based on
the following equations:
Tb = T1 +(T2 − T1)(Hc − H1)
H2 − H1(1)
where Tb is the threshold air temperature at the rain/snow
boundary (◦C); T1 and T2 (◦C) are observed air temperatures
ataltitude H1 and H2 (m), respectively; and Hc is the altitude of
the rain/snow boundary (m).
If the whole HRB is covered in rain, then the interpolated air
temperature at 4765 m was selected as the correspondingair
temperature for the rainfall event, and if the whole HRB is covered
with snow, the air temperature at 2980 m wasselected as the
corresponding air temperature for the snowfall event. The
basin-scale precipitation types and correspondingair temperatures
are shown in Fig. 9c. When rainfall occurred all over the HRB, the
air temperature at 4765 m was greaterthan 1.0 ◦C; snowfall occurred
when the air temperature was below 0.9 ◦C. Observations indicated
that the air temperatureshifted from −0.3 ◦C to 7.4 ◦C at the
rain/snow boundary, which means that the threshold air temperature
at the rain/snowboundary varied.
Through manual precipitation type observation at 2980 m, EOS
600D camera assisted precipitation type discriminationat 4550 m,
and basin-scale precipitation type observation, we conclude that
the threshold air temperature between rain andsnow is generally 0
◦C in the HRB. The rain and mixed precipitation threshold air
temperature is not obvious in the HRB(Fig. 9). The mixed
precipitation air temperature ranges from 0 ◦C to 7.6 ◦C.
4.3. Half-hour air temperatures for the three precipitation
types
Precipitation is usually concentrated at a specific time of day,
which means the precipitation type is determined moreby the air
temperature at these specific times than by the daily resolution
air temperature. Fig. 10 shows the half-hour airtemperatures during
the three different precipitation types, in which the wide bars
represent the upper quartile and thelower quartile of half-hour air
temperature, the plus sign represents the median half-hour air
temperature, and the finebars represent the maximum and minimum
half-hour air temperatures during precipitation events. During rain
events, thehalf-hour air temperatures were above 0 ◦C (Fig. 10a),
the half-hour air temperature at rain/snow boundary were also
above
0 ◦C (Fig. 10b). For mixed precipitation above the rain/snow
boundary, the half-hour air temperature was either range
frompositive to negative or above 0 ◦C (Fig. 10c). For the snow
events, half-hour air temperatures are below 0 ◦C (Fig. 10d).
From these half-hour air temperature analyses, we conclude that
if the half-hour air temperatures are below 0 ◦C, theprecipitation
can be classified as snow; if half-hour air temperatures ranges
from positive to negative during the precipitation
-
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 29
Fig. 11. Relative humidity during three precipitation types: (a)
rainfall, (b) mixed precipitation and (c) snowfall.
ep
4
at
m8hhThHo
vent, the precipitation can be classified as mixed
precipitation; and if the half-hour air temperatures are above 0
◦C, therecipitation type can be either rain or mixed
precipitation.
.4. Relative humidity for rainfall, mixed precipitation, and
snowfall
Ding et al. (2014) and Harder and Pomeroy (2013) have shown that
precipitation types are related to a combination ofir temperature
and humidity. We selected the half-hour relative humidity during
precipitation events at different altitudeso analyze the
relationship of relative humidity and precipitation types.
Half-hour relative humidity observations in the HRB indicated
that relative humidity is lower for snowfall events than forixed
precipitation or rainfall (Fig. 11). Most of the observed relative
humidity values at the rain/snow boundary were over
5%. Photographed weather conditions were cloudy or foggy during
mixed precipitation or rainfall, which indicates veryigh humidity
conditions. Mixed precipitation events were concentrated in the
warm season, which corresponds with highumidity conditions. Some
mixed precipitation events were actually shifts from rainfall to
mixed or rainfall to snowfall.he falling rain adds moisture to the
air, combined with falling air temperatures during precipitation
event, this causesigh humidity conditions. In sub-saturated
conditions, snowflakes will sublimate instead of melt into liquid
water. In theRB, cloudy or foggy weather conditions indicate that
sublimation or evaporation was not common. This explains why
the
bserved threshold air temperature was 0 ◦C for mixed
precipitation and rainfall.
-
30 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
y = 0.994xR² = 0.845
0
25
50
75
100
0 25 50 75 100
)%(
m0433 ta ytidimu
H evitaleR
Relative Humidity at 2980m(%)
a
y = 0.965xR² = 0.773
0
25
50
75
100
0 25 50 75 100
)%(
m0173 ta y ti dimu
H evitaleR
Relative Humidity at 2980m(%)
b
y = 0.917xR² = 0.542
0
25
50
75
100
0 25 50 75 100
)%(
m6614ta ytidimu
H evitaleR
Relative Humidity at 29 80m(%)
c
y = 1.038xR² = 0.432
0
25
50
75
100
0 25 50 75 100
)%(
m6944ta yti dimu
H e vi taleR
Relative Humidity at 29 80m(%)
d
Fig. 12. Comparison of observed relative humidity at four
different locations: (a) between 2980 m and 3340 m; (b) between
2980 m and 3710 m; (c) between2980 m and 4166 m; (d) between 2980 m
and 4496 m.
5. Discussion
5.1. Humidity at different altitudes
Observations in the HRB indicated that the correlation
coefficients of humidity at four different altitudes decreased
withincreasing distance (Fig. 12). A paired samples test indicates
that the standard deviation for relative humidity is greater
than25% between two different sites, and the standard error mean is
more than 16%, and the differences are significant
spatially.Estimation of humidity from other measurements will have
a large uncertainty if the measured sites are not near to
eachother. If the humidity was not available, the calculation of
humidity based on observations from other site observationscould
cause great uncertainty. Use of air temperature and estimated
humidity to discriminate precipitation type could alsoinduce great
uncertainty.
5.2. Liquid proportion in mixed precipitation
Discrimination of rain, snow and mixed precipitation and
quantifying the liquid proportion in mixed precipitation
arecritical to runoff estimations. In this paper, we monitored
precipitation types with time-lapse cameras and manual
observa-tions at point or basin scales, which enabled us to observe
precipitation type at different altitudes in a mountainous
region.Although we applied time-lapse photography to monitor
rain/snow boundaries, and obtained threshold air temperaturesto
differentiate rain, snow, and mixed precipitation, we need to
quantify the liquid proportion in mixed precipitation. Therain/snow
boundary is the upper limit of rain and lower limit of mixed
precipitation. For other mountainous regions withoutbasin-scale
observations of precipitation type, we need to find appropriate
ways to discriminate precipitation types. In thiscase, a comparison
of various precipitation discrimination approaches is needed in the
HRB to evaluate precision at differ-
ent time scales. Future observations need to monitor the liquid
proportion in mixed precipitation, which can improve theprecision
of precipitation discrimination methods especially for mixed
precipitation.
-
5
aEtibi
ttod
ottWp
dstMt
6
saodtttrfl
C
A
S
A
h
R
B
B
D
J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32 31
.3. Threshold air temperature at different time scales
In this paper, we investigated daily resolution and half-hour
threshold air temperatures for rainfall, mixed precipitationnd
snowfall. At daily resolution, threshold air temperatures at two
observation points (manual observation at 2980 m, andOS 600D camera
at 4550 m) and basin-scale observation results are not identical.
We found that rainfall happened if the airemperature was above 0
◦C, but the precipitation type can be classified only unequivocally
as rainfall if the air temperatures above 7.6 ◦C. Snow happened if
the air temperature was below 0 ◦C. The air temperature at the
rain/snow boundary shiftedetween 0 ◦C and 7.6 ◦C, so the
differentiation of rainfall and mixed precipitation through a
static threshold air temperature
s not appropriate.Half-hour air temperatures during the
precipitation event combined with corresponding precipitation type
indicated
hat rainfall happened when the half-hour air temperatures were
above 0 ◦C and snowfall happened when the half-hour airemperatures
were below 0 ◦C. When mixed precipitation happened, the half-hour
air temperatures could be either positiver have crossed 0 ◦C during
the precipitation event. The half-hour air temperatures at
rain/snow boundary were above 0 ◦Curing the precipitation
event.
At either a daily or half-hour resolution, 0 ◦C is the threshold
air temperature between snow and rain. Discriminationf mixed
precipitation and rain based on the threshold air temperature is
not appropriate on a daily scale, because the airemperature at the
rain/snow boundary fluctuated from 0 ◦C to 7.6 ◦C. At a half-hour
resolution, when the half-hour airemperatures during the
precipitation events are above 0 ◦C, the precipitation could either
be rain or mixed precipitation.
hen the half-hour air temperatures fluctuated between negative
and positive values during precipitation events, therecipitation
type could be discriminated as mixed precipitation.
Past precipitation type observations in the Qilian Mountains
usually identified two static threshold air temperatures
toiscriminate snow, mixed and rain precipitation. For example,
observations at the July 1st Glacier indicate that the twotatic
threshold air temperatures are 2.3 ◦C and 7.2 ◦C. Within this
range, the probability of rainfall increased with rising
airemperature from 0 to 100% (Sakai et al., 2006). The two critical
temperatures observed at the Yanglong River in the Qilian
ountains are 0 ◦C and 7.2 ◦C (Ding and Kang 1985). The lower and
upper limits from our observation results are similar tohose of
Ding and Kang (1985), except that in the HRB rainfall is likely to
happen in this threshold temperature range.
. Conclusions
A comparative study of precipitation in two successive years in
the HRB indicates that rising air temperatures inducedhifts from
mixed precipitation to rain at high altitudes. Point and
basin-scale observations indicated that precipitation typest high
altitudes are more sensitive to air temperature changes than
precipitation type at lower altitudes. Precipitation
typebservations in the HRB indicate that at daily or half-hour
resolutions, 0 ◦C can be used to discriminate rain and snow. Ataily
resolution, air temperatures at the rain/snow precipitation
boundary fluctuated between 0 ◦C and 7.6 ◦C, which meanshat
precipitation can be discriminated as rain only when the daily
resolution air temperatures were above 7.6 ◦C. From 0 ◦Co 7.6 ◦C,
the precipitation type could be either rain or mixed precipitation.
During a precipitation event, the half-hour airemperatures were
above 0 ◦C at the rain/snow precipitation boundary, the half-hour
air temperatures were above 0 ◦C whenainfall occurred, the
half-hour air temperatures were below 0 ◦C when snowfall occurred,
and the half-hour air temperatureuctuated around 0 ◦C when mixed
precipitation occurred.
onflict of interest
None.
cknowledgments
This paper is supported by the National Basic Research Program
of China (2013CBA01806) and the National Naturalcience Foundation
of China (91025011, 41401078, 41222001, 91125002).
ppendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version,
atttp://dx.doi.org/10.1016/j.ejrh.2015.11.013.
eferences
ox, J., Fettweis, X., Stroeve, J., Tedesco, M., Hall, D.,
Steffen, K., 2012. Greenland ice sheet albedo feedback:
thermodynamics and atmospheric drivers.
Cryosphere 6, 821–839.
raun, L.N., 1985. Simulation of snowmelt-runoff in lowland and
lower alpine regions of Switzerland. Dissert. Zürcher Geographische
Schriften, Heft 21.(ETH) Zürich.
ing, B., Yang, K., Wang, L., Chen, Y., He, X., 2014. The
dependence of precipitation types on surface elevation and
meteorological conditions and itsparameterization. J. Hydrol. 513,
154–163.
http://dx.doi.org/10.1016/j.ejrh.2015.11.013http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0005http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0015
-
32 J. Liu, R. Chen / Journal of Hydrology: Regional Studies 5
(2016) 20–32
Ding, L.F., Kang, X.C., 1985. Climatic conditions for the
development of glacier and their effect on the characteristics of
glaciers in Qilian Mountain. In:Memoir Lanzhou Institute Glaciology
and Geocryology. Chinese Academy of Sciences No. 5 (in Chinese)
Beijing Science Press, pp. 9–15.
Fassnacht, S.R., Yang, Z.L., Snelgrove, K.R., Soulis, E.D.,
Kouwen, N., 2006. Effects of averaging and separating soil moisture
and temperature in the presenceof snow cover in SVAT and
hydrological model for a southern Ontario, Canada, watershed. J.
Hydrometeorol. 7, 298–304.
Haji, M., 2007. Automated discrimination of precipitation type
using the FD12P present weather sensor:evaluation and
opportunities,KoninklijkNederlandsMeteorologischInstituut. R&D
Information and Observation Technology De
Bilt,http://www.knmi.nl/publications/fulltexts/rp pwsneerslagsoort
knmi 20071025 tr297 pdf, 1–77.
Harder, P., Pomeroy, J., 2013. Estimating precipitation phase
using a psychrometric energy balance method. Hydrol. Processes 27,
1901–1914.Leroy, M., Bellevaux, C., 1998. PREWIC: the WMO
intercomparison on present weather sensors/systems (Canada and
France, 1993–1995). Final report.
WMO, Geneva, Switzerland, IOM73 (TD887).Liu, J.F., Chen, R.S.,
Song, Y.X., 2012. Mapping snowcover based on digital
Photogrammerry. J. Glaciol. Geocryol. 34 (3), 618–624 (in
Chinese).Liu, J.F., Chen, R.S., Song, Y.X., Yang, Y., Qing, W.W.,
Han, C.T., Liu, Z.W., 2014. Observations of precipitation type
using a time-lapse camera in a
mountainous region and calculation of the rain/snow proportion
based on the critical air temperature. Environ. Earth Sci. 73,
1545–1554.Loth, B., Graf, H.F., Oberhuber, J.M., 1993. Snow cover
model for global cimate simulations. J. Geophys. Res. 98,
10451–10464.Reeves, H., Elmore, K., Ryzhkov, A., Schuur, T.,
Krause, J., 2014. Sources of uncertainty in precipitation type
forecasting. Weather Forecasting 29 (4),
936–953.Sakai, A., Matsuda, Y., Fujita, K., 2006. Meteorological
observation at July 1st Glacier in northwest China from 2002 to
2005. Bull. Glaciol. Res. 23, 23–32.Sugiura, K., Yang, D., Ohata,
T., 2003. Systematic error aspects of gauge-measured solid
precipitation in the Arctic, Barrow, Alaska. Geophys. Res. Lett.
30
(4), 1192.Thériault, J.M., Stewart, R.E., Henson, W., 2010. On
the Dependence of winter precipitation types on temperature,
precipitation rate, and associated
features. J. Appl. Meteorol. Climatol. 49, 1429–1442.Yang, D.Q.,
Goodison, E.B., Metcalfe, J.R., Golubev, V.S., Elomaa, E., Gunther,
T., Bates, R., Pangburn, T., Hanson, C.L., Emerson, D., Copaciu,
V., Milkovic, J.,
1995. Accuracy of Tretyakov precipitation gauge: result of WMO
intercomparison. Hydrol. Processes 9, 877–895.
http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0020http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0025http://www.knmi.nl/publications/fulltexts/rp_pwsneerslagsoort_knmi_20071025_tr297http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0035http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0045http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0050http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0055http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0060http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0065http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0070http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0075http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080http://refhub.elsevier.com/S2214-5818(15)00128-7/sbref0080
Discriminating types of precipitation in Qilian Mountains,
Tibetan Plateau1 Introduction2 Study area and data3 Methods4
Results4.1 Precipitation type observation results at three
locations4.2 Daily mean threshold air temperatures in the HRB4.3
Half-hour air temperatures for the three precipitation types4.4
Relative humidity for rainfall, mixed precipitation, and
snowfall
5 Discussion5.1 Humidity at different altitudes5.2 Liquid
proportion in mixed precipitation5.3 Threshold air temperature at
different time scales
6 ConclusionsConflict of interestAcknowledgmentsAppendix A
Supplementary dataReferences