229
A wintertime warming trend (0.15 °C year
−1
) has also been observed for GH
which is well captured by all datasets but NCEP-NCAR captured this trend precisely, i.e. 0.18 °C year
−1
.Wintertime precipitation has also increased (~8.58 mm/
year) over GH in past 25 years and this rise has been captured by all datasets but
ERA-I (~4.29 mm/year) and GPCC (3.71 mm/year) performed better than others.
A warming trend (0.08 °C year
−1
) in wintertime temperature was reported for
KH which was more closely captured by ERA-I (0.07 °C/year) and UDEL (0.06 °C/
year). During same time,
declining wintertime precipitation trends (−4.57 mm/year) were observed for
KH. This trend was captured by UDEL, ERA-I and CRU-TS.
The two-step evaluation of datasets revealed efficiency of datasets at different
zones. Since the fidelity of gridded datasets depends mainly upon its spatial resolution, we omitted the use of datasets with coarser resolution, i.e. NCEP-NCAR and
GPCP for further analyses. Also, few datasets which do not provide temperature
data like TRMM and GPCC were also excluded. APHRODITE because of its ceased
temporal coverage after year 2007 was also excluded. Thus, only two datasets:
ERA-I and CRU-TS were employed to appraise spatio-temporal variations of
annual temperature and precipitation over NWH.
4.2 Spatial Variability in Climate as Captured by Selected
Datasets
4.2.1 Annual Mean Temperature
As explained earlier that NWH has a wide altitudinal range due to which temperature distribution has substantial variability spatially. As per observations, annual
average temperature over LH, GH and KH is approximately 9.6 °C, 3.2 °C and
−8.1 °C respectively. Thus we can observe that average temperature declines while
traversing from LH to KH which could be seen as the consequence of altitude. This
distribution is well captured by both ERA-I and CRU-TS (Fig. 6). ERA-I shows a
bit of cold bias by estimating annual Tmean approx. 6.8 °C, −2.5 °C and −8.2 °C
for LH, GH and KH respectively. However, the reported bias was minimal for
KH. Unlike ERA-I, CRU-TS showed warm bias in estimating annual Tmean approx.
14.9 °C, 1.2 °C and −2.7 °C for LH, GH and KH respectively. These biases are
inherent in any dataset owing to generation methods. CRU-TS is an interpolated
dataset and since not much observatories are included from elevations above 5000 m
while interpolation, the estimated values reflect the predominance of data from low
altitude observatories. The observed biases also suggest that bias correction of these
datasets is imperative before their use in hydrological applications.
An Appraisal of Spatio-Temporal Characteristics of Temperature and Precipitation…
A wintertime warming trend (0.15 °C year
−1
) has also been observed for GH
which is well captured by all datasets but NCEP-NCAR captured this trend precisely, i.e. 0.18 °C year
−1
.Wintertime precipitation has also increased (~8.58 mm/
year) over GH in past 25 years and this rise has been captured by all datasets but
ERA-I (~4.29 mm/year) and GPCC (3.71 mm/year) performed better than others.
A warming trend (0.08 °C year
−1
) in wintertime temperature was reported for
KH which was more closely captured by ERA-I (0.07 °C/year) and UDEL (0.06 °C/
year). During same time,
declining wintertime precipitation trends (−4.57 mm/year) were observed for
KH. This trend was captured by UDEL, ERA-I and CRU-TS.
The two-step evaluation of datasets revealed efficiency of datasets at different
zones. Since the fidelity of gridded datasets depends mainly upon its spatial resolution, we omitted the use of datasets with coarser resolution, i.e. NCEP-NCAR and
GPCP for further analyses. Also, few datasets which do not provide temperature
data like TRMM and GPCC were also excluded. APHRODITE because of its ceased
temporal coverage after year 2007 was also excluded. Thus, only two datasets:
ERA-I and CRU-TS were employed to appraise spatio-temporal variations of
annual temperature and precipitation over NWH.
4.2 Spatial Variability in Climate as Captured by Selected
Datasets
4.2.1 Annual Mean Temperature
As explained earlier that NWH has a wide altitudinal range due to which temperature distribution has substantial variability spatially. As per observations, annual
average temperature over LH, GH and KH is approximately 9.6 °C, 3.2 °C and
−8.1 °C respectively. Thus we can observe that average temperature declines while
traversing from LH to KH which could be seen as the consequence of altitude. This
distribution is well captured by both ERA-I and CRU-TS (Fig. 6). ERA-I shows a
bit of cold bias by estimating annual Tmean approx. 6.8 °C, −2.5 °C and −8.2 °C
for LH, GH and KH respectively. However, the reported bias was minimal for
KH. Unlike ERA-I, CRU-TS showed warm bias in estimating annual Tmean approx.
14.9 °C, 1.2 °C and −2.7 °C for LH, GH and KH respectively. These biases are
inherent in any dataset owing to generation methods. CRU-TS is an interpolated
dataset and since not much observatories are included from elevations above 5000 m
while interpolation, the estimated values reflect the predominance of data from low
altitude observatories. The observed biases also suggest that bias correction of these
datasets is imperative before their use in hydrological applications.
An Appraisal of Spatio-Temporal Characteristics of Temperature and Precipitation…