Pure Appl. Geophys. Ó 2013 Springer Basel DOI 10.1007/s00024-013-0675-9
Pure and Applied Geophysics
Role of the Himalayan Orography in Simulation of the Indian Summer Monsoon using RegCM3 P. SINHA,1 U. C. MOHANTY,1 S. C. KAR,2 and S. KUMARI1 Abstract—In this study, sensitivity of the Indian summer monsoon simulation to the Himalayan orography representation in a regional climate model (RegCM) is examined. The prescribed height of the Himalayan orography is less in the RegCM model than the actual height of the Himalayas. Therefore, in order to understand the impact of the Himalayan orography representation on the Indian summer monsoon, the height of the Himalayan orography is increased (decreased) by 10 % from its control height in the RegCM model. Three distinct monsoon years such as deficit (1987), excess (1988) and normal rainfall years are considered for this study. The performance of the RegCM model is tested with the use of a driving force from the reanalysis data and a global model output. IMD gridded rainfall and the reanalysis-2 data are used as verification analysis to validate the model results. The RegCM model has the potential to represent mean rainfall distribution over India as well as the upper air circulation patterns and some of the semi-permanent features during the Indian summer monsoon season. The skill of RegCM is reasonable in representing the variation in circulation and precipitation pattern and intensity during two contrasting rainfall years. The simulated seasonal mean rainfall over many parts of India especially, the foothills of the Himalaya, west coast of India and over the north east India along with the whole of India are more when the orography height is increased. The low level southwesterly wind including the Somali jet stream as well as upper air circulation associated with the tropical easterly jet stream become stronger with the enhancement of the Himalayan orography. Statistical analysis suggests that the distribution and intensity of rainfall is represented better with the increased orography of RegCM by 10 % from its control height. Thus, representation of the Himalayan orography in the model is close to actual and may enhance the skill in seasonal scale simulation of the Indian summer monsoon. Key words: Indian summer monsoon, RegCM, seasonal simulation, Himalayan orography.
1 Centre for Atmospheric Sciences, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India. E-mail:
[email protected] 2 National Centre for Medium Range Weather Forecasting, Noida, India.
1. Introduction Complex topographical features and land surface characteristics play an important role in modulating the local weather and climate (GIORGI and MEARNS 1991; GIORGI and AVISSAR 1997 ; PIELKE 2001), thus land surface and topography have a large influence in regional and local climates (PIELKE and AVISSAR 1990; DICKINSON 1995). KASAHARA (1980) hasprovided a theoretical overview of the influence of orography on atmospheric circulation by considering the dynamical and thermal factors individually and then their combined effect. COOK and HELD (1992) have examined the stationary waves generated over orography using a general circulation model and a linear model. The Indian summer monsoon (ISM) plays an important role for the agro economic country like India as more than 70 % of annual rainfall occurs during summer monsoon seasons (June to September; JJAS) over India (PARTHASARATHY et al. 1994). The prediction of the Indian summer monsoon rainfall (ISMR) in monthly to seasonal scale is crucial for policy planning and decision making. However, the prediction of ISMR is one of the challenging issues to the meteorologists to date. Research on the role of the Himalayan mountain/orography upon the ISM circulation and associated rainfall was initiated about a century ago (BANERJEE 1929). Previous studies using a primitive equation model suggested that the low level monsoon circulation is sensitive to the mountain barrier around that region (DAS and BEDI 1978; GROSSMAN and DURAN 1984). With the advancement of the general circulation models (GCMs), several number of modeling studies have been conducted to examine the role of orography in simulating the south Asian and Indian summer monsoon (HAHN and MANABE 1975; DAS and BEDI 1978; KRISHNAMURTI
P. Sinha et al.
Pure Appl. Geophys.
et al. 1984; ROY ABRAHAM et al. 1996; CHAKRABORTY et al. 2002; ABE et al. 2003). BOOS and KUANG (2010) have shown that the Himalayan mountains produce a strong monsoon by insulating warm, moist air over continental India from the cold and dry extra-tropics. A number of sensitivity studies have been carried out to examine the impact of different land surface properties in simulating the Asian and Indian summer monsoon using GCMs or simple models. FENNESSY et al. (1994) conducted various experiments to study the sensitivity of the ISM simulation by changing different parameters such as orography, vegetation, soil wetness, and cloudiness. They concluded that the representation of orography plays the most important role in simulation of the Indian monsoon circulation and rainfall. HAHN and MANABE (1975) and CHAKRABORTY et al. (2002) have investigated the importance of mountains in simulation of onset and temporal variations of the ISM. ZHONGFENG et al. (2010) examined the role of different mountains using atmospheric GCMs in simulation of the Asian summer monsoon. Their results illustrate that the presence of Asian mountains produces a stronger monsoon with enhanced lower-tropospheric westerly winds, uppertropospheric easterly winds, and stronger water vapor convergence. However, most of these studies are hypothetical in the sense of assuming with and without orography conditions and do not emphasize about the changes of the summer monsoon due to what degree of changes are there for mountain height. Efforts have been made by using ‘‘enveloped orography’’ or considering different mountain regions by increasing the height linearly in atmospheric or coupled GCMs to study the impact of orography in simulation of the Asian as well as the Indian summer monsoon. KRISHNAMURTI et al. (1984) carried out experiments using enveloped orography and suggested that the track of onset vortex over the Arabian Sea associated with the ISM is represented better with steeper orography. ROY ABRAHAM et al. (1996 have found that the large scale flow associated with the ISM is closer to the observations when the mountain height is enhanced by two standard deviations from its mean. PRELL and KUTZBACH 1992) found an increase of monsoon rainfall with the increase of the Himalayan orography. AN et al. (2001) examined the evolution of Asian monsoon with the uplifting of the Himalaya and found stronger
monsoon activities with the enhancement of the Himalayan mountain region. ABE et al. (2003) examined the changes of the Asian summer monsoon with progressive mountain uplift using coupled GCMs. They found that the ISM becomes stronger gradually with mountain uplift; however, this enhancement is more when the increase of height is made over steeper mountain regions. MOHANDAS et al. (2009) have conducted a number of sensitivity experiments to find out the possible impacts of filtered modified orography using spectral GCM in prediction of ISM. Their results indicate that enhancement of the peaks of the Himalayan range and Western Ghats in the GCM are able to represent better circulation features and associated rainfall during the ISM. Since the representation of the land surface characteristics will be better in regional climate models (RCMs) than a GCM, thus, it will be interesting to study the sensitivity of the Himalayan orography representation using RCM in simulating ISM. The impact of the Tibetan Plateau (TP) upon the Asian summer monsoon is studied using the National Center for Environment Prediction (NCEP)-regional spectral model (SONG et al. 2010). They studied the variations in monsoon activities with the progressive uplifting of TP and found that simulated precipitation is more over many parts of India such as the west coast, and north India with the increase of TP. So far, sensitivity of the Himalayan orography representation using the regional climate model RegCM (version 3.1) has not been studied in the simulation of ISM. SINHA et al. (2012a) have carried out an extensive study on use of RegCM forced with observed analysis and outputs from a GCM for three contrasting monsoon years. They have carried out nested grids at 90 and 30 km horizontal resolutions with various convective parameterization schemes. They found that the model is able to bring out essential features of the mean monsoon circulation and rainfall reasonably well. It is not clear how the Himalayan orography responds to the circulation anomalies during contrasting monsoon years (with weak, normal and strong monsoon winds) and how the rainfall patterns over India gets modified with the specified heights of the Himalayan mountains in the regional climate model RegCM. In this study, the impact of the Himalayan orography representation in simulating the ISM using RegCM (version 3.1) has been examined for deficit,
Role of the Himalayan Orography
excess and normal monsoon rainfall years. For this purpose, two extreme rainfall years in which one is excess (1988) and another is deficit (1987) rainfall year, and a normal monsoon year of 1989 (SINHA et al. 2012a) are considered. The RegCM model has been forced by the NCEP—Department of Energy (DOE) reanalysis 2 (KANAMITSU et al. 2002) and outputs from a GCM in order to understand the role of different heights of the Himalayan region in simulation of ISM. A brief description of the models and data used in this study has been described in Sect. 2. Experimental design and methodology have been represented in Sect. 3. Analyses of the results and associated discussions have been provided in Sect. 4. Finally, conclusions of this study have been given in Sect. 5.
limited-area model. The dynamical component of the hydrostatic version of the MM5 model (GRELL et al. 1994) is used in the RegCM. The model is a compressible and finite difference method used to solve the equations. Similar to MM5, the RegCM model has the terrain following vertical r-coordinates. The model takes account of various physical exchanges between the land and atmosphere through several parameterizations assumptions. The standard vertical levels in RegCM are 18 with five levels within the planetary boundary layer (up to around 850 hPa). Several number of cumulus schemes are included in the model. A number of lateral boundary conditions are present for treatment in the interface of large scale GCMs output or reanalysis and the RegCM model. The details of the model description can be found in PAL et al. (2007).
2. Model and Data Used The models used in this study are the National Centre for Medium Range Weather Forecasting (NCMRWF) global spectral model and International Centre for Theoretical Physics (ICTP) regional climate model RegCM (version 3.1). Brief descriptions of these models are provided below. 2.1. NCMRWF Model The global model used for this study is the Indian global model (KAR 2007) which is the climate version of forecast model of NCMRWF. This global model is spectral in nature having 80 waves in Triangular truncation (T80). The horizontal resolution of this model is equivalent to 150 km in both zonal and meridional directions. The model vertical coordinate is sigma and consists of 18 vertical layers. In the NCMRWF model, Kuo-Anthes type of scheme is used for deep convection (ANTHES 1977). In each year, the model is initialized in May and integrated until the end of September. More details of the model may be found at KAR (2007), KAR et al. (2011) and SINHA et al. (2012a, b). 2.2. RegCM Model The regional climate model RegCM (PAL et al. 2007) (version 3.1), developed at the International Centre for Theoretical Physics (ICTP), Italy, is a
2.3. Data Used The NCMRWF model has been initialized with reanalysis data and sea surface temperature (SST) is provided from NCEP–Climate Forecast Systems (CFS) forecasted SST (version-1). The NCEP-CFS forecasted SST is available in the site http://cfs.ncep. noaa.gov/. The regional climate model is driven by both NCEPDOE reanalysis 2 (referred to as NNRP2 hereafter) data at 2.5° resolution (KANAMITSU et al. 2002) and NCMRWF model output at 1.4° horizontal resolution (KAR et al. 2011) separately. The weekly mean SST from the National Oceanic and Atmospheric Administration at 1° 9 1° and other geophysical parameters from the United State Geological Survey at 30 min resolution are used to provide surface boundary conditions. Observed daily rainfall data at 1° 9 1° resolution (RAJEEVAN et al. 2006) obtained from the India Meteorological Department (IMD) and upper air meteorological data from NNRP2 have been used for verification of the RegCM model products.
3. Experimental Design and Methodology In this study, NCMRWF-GCM is initialized in May and integrated until the end of September each year separately to generate the initial and boundary
P. Sinha et al.
conditions for the regional climate model RegCM. In each simulation, the RegCM model is initialized on 1st May and integrated up to 30th September separately for each year. In order to study the role of different Himalayan orography representation in RegCM to simulate ISM, two sets of experiments have been designed. Set-1: RegCM driven by NNRP2 data 1. CONTR with mean orography height over Himalayas. 2. EHO1R—Himalayan orography reduced by 10 %. 3. EHO2R—Himalayan orography enhanced by 10 %. Set-2: RegCM driven by GCM data
Pure Appl. Geophys.
Table 1 RegCM3 model configuration used in the present study Dynamics
Hydrostatic
Main prognostic variables Model domain
u, v, t, q and p0
Map projection Vertical co-ordinate Lateral boundary conditions Cumulus parameterization Radiation parameterization PBL parameterization
30°S–60°N; 28°E–128°E; Res. = 90 km ROTMER Terrain-following sigma co-ordinate. Total 18 sigma levels. Exponential relaxation Grell cumulus with Arakawa Schubert closer NCAR/CCM3 radiation scheme Holtslag
1. CONTM with mean orography height over Himalayas. 2. EHO1M—Himalayan orography reduced by 10 %. 3. EHO2M—Himalayan orography enhanced by 10 %.
shown in Fig. 1. In this study, seasonal simulation of the Indian summer monsoon has been carried out for three distinct monsoon rainfall years viz. 1987 (deficit), 1988 (excess) and, 1989 (normal) years.
The initial and boundary conditions have been provided from 6-hourly NNRP2 data and the NCMRWF global model forecasted data at 12 hourly intervals. The forcing data are interpolated spatially to the model grid and temporally to the model time step. In this study, the Himalayan orography is increased (decreased) in the RegCM from its control height as follows:
4. Results and Discussion
H ¼ a H ð H 1:5 kmÞ;
ð1Þ
where H is the modified height, H is the control height in RegCM. The value of a is 1.1 (0.9) for 10 % increase (decrease) of the Himalayan height from control height in RegCM. The value of a is 1 for H\1:5 km. In the first set of experiments, evaluation of performance of the RegCM forced by NNRP2 data has been carried out. In this set, a number of sensitivity experiments using three different Himalayan orographies have been carried out. In the second set, sensitivity of the Indian summer monsoon simulation to the Himalayan orography representation in RegCM forced by GCM has been examined. The model configuration is presented in Table 1 and model domain is
The results presented in this study are analyzed in three broad sections, namely, (1) prevailing synoptic situations, (2) RegCM driven by NNRP2 data and, (3) RegCM driven by GCM data. The detailed discussions and analyses are given below: 4.1. Prevailing Synoptic Situations The prevailing synoptic situations that occurred during two contrasting and one normal monsoon rainfall years are here described by rainfall maps from IMD data and wind pattern and magnitude at 850 hPa as well as 200 hPa from NNRP2 data. Seasonal (JJAS) mean of rainfall obtained from IMD and wind at 850 hPa as well as at 200 hPa obtained from NNRP2 data for the year 1987, 1988 and 1989 are shown in panels a–i of Fig. 2, respectively. The variations in rainfall and circulation patterns among three distinct years have been observed by computing the difference between excess, deficit and normal years. The differences between excess and deficit years (1988–1987) as well as excess and normal years (1988–1989) are
Role of the Himalayan Orography
Figure 1 RegCM3 model domain (at 90 km horizontal resolution) used for the present study
computed for rainfall and winds at 850 and 200 hPa. These are shown in panels a–f of Fig. 3, respectively. Figure 2a, d, g represent the seasonal rainfall during monsoon season over India for 1987 (deficit), 1988 (excess) and, 1989 (normal) years, respectively. Observations indicate that seasonal mean rainfall intensity is maximum over the west coast and north east India and minimum over northwest India, while the intensity is moderate over the remaining parts of the Indian region during all the years. However, the magnitude of seasonal rainfall varies region to region and season to season. The occurrence of less rainfall is observed over many parts of India such as Jammu and Kashmir, northwest India, west coast of India, southern peninsular India, and central India in 1987 than 1989. It is also noticed that the observed rainfall over these regions in 1989 is comparatively less than 1988. Seasonal rainfall over major parts of north east India and some parts of east India are more in 1987 as compared to 1988 and 1989. The occurrence of more rainfall over the north-east part of India during the deficit year is mainly attributed to eastward shifting
of Hadley circulation and the ascending motion of this circulation confined over this region. Analysis of lower level (at 850 hPa pressure level) circulation shows that the seasonal mean wind over the Arabian Sea (AS) including the Somali Jet (SJ) has more strength (around 20 ms-1) in 1988 than 1987 (around 15 ms-1). It is noted that the strength of SJ is more in 1989 (of about 17 ms-1) than 1987 and less than 1988. It is also observed that the strength of the cross equatorial flow is lesser in 1987 and higher in 1988 as compared to 1989. The variations in low level circulations among the three different years illustrate that southwesterly flow over the AS and northerly/ northeasterly flow over the Bay of Bengal (BoB) are stronger near the Indian coast in excess than in deficit and normal years. As a result, the stronger wind brings more moisture from AS and BoB to the Indian landmass causing more rainfall over many parts of India during an excess rainfall year. Observational analysis of upper air seasonal mean wind at 200 hPa indicates that the axis of the Tropical Easterly Jet (TEJ) persists around 9°N latitude during all the years.
P. Sinha et al.
Pure Appl. Geophys.
Figure 2 Seasonal (June to September, JJAS) mean of a IMD gridded rainfall (mm/day), b NCEP-DOE reanalysis 2 wind (m/s) at 850 and, c NCEPDOE reanalysis 2 wind (m/s) at 200 hPa for 1987; d, e and, f are same as a, b and, c, respectively, but for 1988; g, h and, i are same as a, b and, c, respectively, but for 1989
However, TEJ exists between 55°E–80°E, 45°E–85°E, and 55°E–80°E in the west-east direction during deficit, excess and normal years, respectively, and the area with stronger TEJ is more during an excess year than normal and deficit years. The subtropical Westerly Jet Stream (WJS) is shifted northwards and located to the north of the Himalaya during monsoon season in all the years; however WJS is weaker in excess years and stronger in deficit years. The variations in upper air circulations indicate that the strength of the TEJ is less in a deficit year than a normal year and following less than an excess year. The seasonal mean sea level pressure (MSLP) and surface air temperature at 2 m height (T2M) obtained from the NNRP2 data for all the years are analyzed (Figures not shown). It is seen that a low pressure area exists over northwest India and
its nearby region during all the years. Mascarene high and higher pressure over the adjoining Indian Ocean persist during all the three years. However, the MSLP over northwest India and its surrounding is minimum (*999 to 1,000 hPa) during excess and normal rainfall years than deficit year (*1,002 hPa). The area with minimum MSLP is larger and the pressure gradient along the west coast of India is stronger during the excess year than deficit and normal years. This result agrees well with the previous study by SINGH et al. (2011) based on a 57 years NCEP reanalysis data that all-India summer monsoon rainfall is positively (negatively) correlated with low (high) MSLP over the Indian region especially over north west India. The kinks in the isobars associated with the orography on the leeward side are seen during all the years.
Role of the Himalayan Orography
Figure 3 Seasonal mean difference of a IMD gridded rainfall (mm/day), b NCEP-DOE reanalysis 2 wind at 850 hPa (m/s), c NCEP-DOE reanalysis 2 wind at 200 hPa (m/s) calculated between excess (1988) and deficit (1987) year; d, e and, f are same as a, b and, c, respectively, but calculated between excess (1988) and normal (1989) year
The T2M over northwest India and its adjoining region is more than the rest of the Indian subcontinent in the corresponding years. When compared with different monsoon years, it is noticed that T2M over the northwest India is more (308–310 K) by an about 1–2 °C during an excess year than a normal year. 4.2. RegCM Driven by NNRP2 Data Firstly, the RegCM model is forced by NNRP2 to examine the performance of the model in simulating monsoon circulation and associated precipitation over
the Indian region. For this purpose, all the three distinct years are considered to carry out the seasonal scale simulations of a summer monsoon. The detail of the model configuration is provided in Table 1. Seasonal mean precipitation and circulations at 850 and 200 hPa over the Indian monsoon domain obtained from RegCM simulations are presented in Fig. 4. The higher precipitation over the west coast and over some parts of northeast India and lesser precipitation over northwest India is represented well in the model simulations during all the years; however the intensity of the model simulated rainfall is less as
P. Sinha et al.
Pure Appl. Geophys.
Figure 4 NCEP-DOE reanalysis 2 driven RegCM3 (at 90 km resolution) simulated seasonal (JJAS) mean of a rainfall (mm/day), b wind (m/s) at 850 hPa and, c wind (m/s) at 200 for 1987; d, e and, f are same as a, b and, c, respectively, but for 1988; g, h and, i are same as a, b and, c, respectively, but for 1989
compared to observations. In the circulation pattern, a stronger southwesterly at the Somali coast at low level and the TEJ as well as the subtropical WJS at upper air are well brought out by the model during all the years; however, the strength of these semi-permanent features of summer monsoon are weaker in the model than verification analysis. It is noted that the locations of these features are depicted well as compared with the verification analysis. It is noticed from the results that the RegCM model at coarse resolution has a weakness in representing the monsoon trough. The monsoon trough, one of the important semi-permanent features, controls the precipitation amount over central and eastern India during summer monsoon. Therefore,
a poor representation of monsoon trough in the RegCM model at 90 km resolution is responsible for very less precipitation over these regions. In order to understand further the poor performance of RegCM model over central and eastern India, seasonal mean vertical pressure velocity (omega) at 500 hPa has been computed for both NNRP2 and RegCM simulations for all the three years and shown in Fig. 5. It is seen from the figure that the omega is positive over north west India and negative at remaining parts of India in the verification analysis during all three years. The RegCM model is able to bring out positive omega over north west India and negative omega over the west coast, some parts of north east and peninsular India
Role of the Himalayan Orography
Figure 5 Seasonal mean pressure vertical velocity (Omega, Pa/s) at 500 hPa obtained from a NCEP-DOE reanalysis 2 and d NNRP2 driven RegCM3 (at 90 km resolution) simulation computed for the year 1987; b and e are same as a and d, respectively, but computed for 1988; c and f are same as a and d, respectively, but computed for 1989
and foot hills of the Himalaya. This feature is well in agreement with the verification analysis. It is noticed from the Fig. 5 that the omega is positive over most parts of central and east India which is in contradiction with the verification analysis in all the three years. This may be one major reason for poor performance of the RegCM model in simulating seasonal rainfall over these regions in all the years. Examination of RegCM simulated MSLP and T2M for all the three years (Figures are not shown) illustrated that the model is capable of bringing out Mascerene high pressure over the Indian Ocean and low pressure over northwest India and adjoining regions during all the years.
The intensity of the high pressure is reasonably well represented, while low pressure is under estimated as compared with the verification analysis. It is observed that the model simulated intensity of low pressure is higher during excess and normal years than deficit years and the area with minimum MSLP is larger during an excess year and these features are well in agreement with the verification analysis. The area with higher T2M over the Bay of Bengal is more during the deficit year that agrees well with the verification analysis. In order to understand the model response in variation of circulation and precipitation patterns and
P. Sinha et al.
Pure Appl. Geophys.
Figure 6 NCEP-DOE reanalysis 2 driven RegCM3 (at 90 km resolution) simulated seasonal mean difference of a rainfall (mm/day), b wind at 850 hPa (m/s), c wind at 200 hPa (m/s) calculated between excess (1988) and deficit (1987) year; d, e and, f are same as a, b and, c, respectively, but calculated between excess (1988) and normal (1989) year
magnitude for three different years, differences between excess and deficit (1988–1987) as well as excess and normal year (1988–1989) of precipitation and circulation (low level as well as upper air) are computed and shown in Fig. 6. It is seen that the
model simulated rainfall is higher over the west coast of India during an excess year than the other two years which is well in agreement with the observations; however it fails to represent more rainfall over northeast India during a deficit year.
Role of the Himalayan Orography
As stated earlier that the RegCM model performance at 90 km resolution is poor in representing the ‘monsoon trough’ and as a result, the model simulates very less precipitation over central and east India. The weakness and strength of the RegCM model at 90 km resolution in the simulation of ISM are described in detail in SINHA et al. (2012a). The difference in circulation pattern illustrates that the low level westerly/southwesterly over eastern parts of AS and upper air TEJ near the equator are stronger during an excess year and agrees well with verification analysis; Therefore, it is seen that the RegCM model has potential in representing mean rainfall and circulations pattern of the summer monsoon. Since our main focus to study the sensitivity of the Himalayan orography upon the Indian summer monsoon, therefore, RegCM can be a useful regional climate model to carry out experiments. 4.2.1 Sensitivity of the Himalayan Orography Representation It is already demonstrated in the previous section that the performance of RegCM is satisfactory in simulating the ISM and RegCM can be used to investigate the impact of the Himalayan orography on the ISM. For this purpose, the height of the Himalayan region has been increased or decreased by 10 % from its mean height (control experiment) in the RegCM. The difference of height between increased (or decreased) and mean orography in RegCM over various regions of the Himalayan mountains is represented in Fig. 7. The mean height of the Himalayan region (control) is shown in Fig. 1. It is seen that the increase (decrease) is more in the steeper orography regions. The results obtained from the RegCM model driven by NNRP2 data for both EHO2R (increase of the Himalayan height by 10 %) and EHO1R (decrease of the Himalayan height by 10 %) experiments are discussed below: The RegCM model mean height and simulated seasonal mean precipitation averaged over the area covered by 75°E–100°E in the west-east and 22°N– 32°N in the north south have been computed for all the years (as well as mean precipitation of all years) and shown in Fig. 8. Interestingly, it is noticed that as the height of the Himalayan orography increases, the
Figure 7 Difference in height (in meters) of the Himalayan mountain region for both addition and subtraction from its mean (control) height in RegCM3 considered for the study. Contour interval is 50 m
Figure 8 RegCM3 simulated area averaged (75°E–100°E; 22°N–32°N) seasonal (JJAS) mean rainfall computed for different mean height of the Himalayan mountain represented in RegCM using m10 (averaged height is 2,464.3 m), control (averaged height is 2,720.5 m) and, p10 (averaged height is 2,976.7 m)
P. Sinha et al.
Pure Appl. Geophys.
area averaged simulated precipitation over the chosen domain also increases. However, the rate of increase of simulated precipitation varies from year to year. It is noticed that mean simulated precipitation increases (considering all years) around 13 % with the increase of orography from control height by 10 % (EHO2RControl), while the increase of simulated precipitation of about 7 % (Control-EHO1R) when the orography increases from EHO1R to control height. This result agrees well with ABE et al. (2003) that the enhancement of rate of increase of simulated precipitation is more when the increase of height is made over steeper mountain region. It is also noticed that the increase in precipitation simulated by the model with increase in Himalayan orography (EH01RControl-EH02R) is nearly linear in 1987 and 1989 years. However, the enhancement in the precipitation amount is high for steeper orography representation during the stronger monsoon season. This is probably due to the strength of the low level southwesterly wind which is higher during excess years and, thus, the moisture that is brought out by the low level wind is more to the Indian landmass. Over the Bay of Bengal, this south westerly wind becomes south easterly and moves towards the Himalayas. The orographical uplifting is more for the steeper orography and; hence, it produces more rainfall near the foothills of the Himalaya during excess years. In order to understand the model response to different heights of Himalayan orography, differences of seasonal mean rainfall, wind at 850 hPa and wind at 200 hPa between EHO2R and EHO1R experiments of RegCM for the years 1987, 1988 and 1989 are computed and shown in Fig. 9. It is seen that the simulated rainfall over the foothills of the Himalayan region and many parts of the northeast regions over India is more in the EHO2R than EHO1R experiments of RegCM during all the years. This may be attributed to the enhancement of the orographical lifting of the air mass and as a result, occurrence of stronger water vapour convergence over these regions (SONG et al. 2010; ZHONGFENG et al. 2010). During 1988 (an excess year) the simulated seasonal rainfall is more in the EHO2R than EHO1R over the west coast of India. Except near the central part of west coast, the simulated rainfall is more in the EHO2R
than EHO1R even for 1987 (a deficit year), However, for 1989, such enhancement in rainfall is not noticed when orography height over the Himalayas is enhanced. Such differences in response to increase in orography may occur due to difference in flow pattern in these years. It is noticed from the Fig. 9b, e, h that the low level (at 850 hPa) southwesterly wind over the Arabian Sea is stronger in the EHO2R experiments. However, larger areas that cover all parts of India and most parts of the Arabian Sea with a stronger southwesterly have been observed during the excess year for the EHO2R experiment. It can be noted here that due to increase in the height of the Himalayan orography, the divergence of upper air uplifted and the air has more potential energy; when this upper air motion occurs from land to ocean and air sinks to complete its cycle, the potential energy converted into kinetic energy, which in turn, increases the strength of the low level circulation. Thus, increase of the Himalayan orography results in the increase/ strengthening of the southwesterly flow over the Arabian Sea. It is noticed in the difference of 850 hPa wind between EHO2R and EHO1R that a cyclonic circulation centered near the Gujarat regions exists during all the years. It is also seen that the wind is stronger at the south of the cyclonic circulations indicating more moisture entrainment from AS to the Indian landmass through the northern part of west coast. During deficit and normal years, the stronger westerly is found over north of the Arabian Sea and northern India in the enhanced Himalayan mountain experiments. Therefore, the low level southwesterly wind associated with the SJ is stronger due to the increase of Himalayan orography. Analysis of the upper air circulation reveals that the TEJ is stronger in the EHO2R experiment than EHO1R during an excess year. However, no notable changes are found in the strength of TEJ in EHO2R and EHO1R experiments during normal and deficit years. It is noticed that a subtropical westerly jet stream that shifted towards north and centered nearly 37°N and 100°E is stronger for EHO2R than EHO1R experiments (Fig. 9c, f, i). These results are well in agreement with the previous studies (ROY ABRAHAM et al. 1996; ABE et al. 2003; MOHANDAS et al. 2009; ZHONGFENG et al. 2010) that the large scale fields such
Role of the Himalayan Orography
Figure 9 NCEP-DOE reanalysis 2 driven RegCM3 (at 90 km resolution) simulated seasonal mean difference between experiments EHO2R (10 % increase of the Himalayan height) and EHO1R (10 % decrease of the Himalayan height) of a rainfall (mm/day), b wind at 850 hPa (m/s), c wind at 200 hPa (m/s) calculated for the year 1987; d, e and, f are same as a, b and, c, respectively, but calculated for 1988; g, h and, i are same as a, b and, c but calculated for 1989
as low level southwesterly, upper air TEJ associated with the ISM are stronger with the increased orography of the Himalayan mountains. 4.3. RegCM Driven by GCM Data It is established in the previous section that the RegCM model has reasonable skill in reproducing the observed features of monsoons when observed analysis is provided as model initial and lateral boundary conditions. It is also illustrated that the performance of the reanalysis driven RegCM model varies with the variation in the representation of the Himalayan orography, in simulating the ISM. Here, the performance of the RegCM model with different orography
representations is examined when forced by a global model simulation. For this, the NCMRWF model seasonal prediction data available at 12 hourly intervals is used. The configuration of the RegCM model is provided in Table 1. The control (hereafter CONTM), EHO2M (with 10 % increased of the Himalayan height) and EHO1M (with 10 % decrease of the Himalayan height) experiments are the same as previous experiments CONTR, EHO2R and, EHO1R, respectively, except the initial and boundary conditions are provided from NCMRWF model output instead of NNRP2 data. The seasonal mean rainfall obtained from the NCMRWF model, CONTM run and, rainfall differences between EHO2M and EHO1M experiments of
P. Sinha et al.
Pure Appl. Geophys.
Figure 10 Seasonal mean rainfall (mm/day) a from NCMRWF model b from NCMRWF model driven RegCM3 control run c difference (mm/day) between experiments EHO2M (10 % increase of the Himalayan height) and EHO1M (10 % decrease of the Himalayan height) calculated for the year 1987; d, e and, f are same as a, b and, c, respectively, but calculated for 1988; g, h and, i are same as a, b and, c but calculated for 1989
RegCM for the year 1987, 1988 and, 1989 are shown in Fig. 10. It is observed from the figure that the NCMRWF simulation shows more rainfall over CNEI and SPI in a deficit year. It is also observed that RegCM simulated less rainfall over central India in all the three years. This is mainly due to poor representation of the monsoon trough as well as pressure vertical velocity (omega) over central and east India. However, simulated rainfall intensity over almost all parts of the west coast is more in RegCM than the NCMRWF model and also closer to
observation. The RegCM model (CONTM) is able to simulate higher precipitation over many parts of northeast India and over some parts of north India near the Gangetic plains than the global model during all the years. It is noticed that simulated rainfall over central India is less in RegCM than the NCMRWF model. It can be mentioned here that the NCMRWF model has several deficiencies in representing the circulation features during excess as well as deficit rainfall years (SINHA et al. 2012b) and it is noticed that the NCMRWF model is not able to represent the
Role of the Himalayan Orography
inflow of the moisture through the east coast of India during an excess year which is in contradiction when compared with the reanalysis. Probably this is one of the reasons for lesser precipitation simulation over central India. Thus, representation of seasonal rainfall distribution is better in the regional climate model than driven GCM and also supported well with previous studies (BHASKARAN et al. 1996; JI and VERNEKAR 1997). In the experiments with different heights of the Himalayan mountains, the seasonal rainfall
differences between EHO2M and EHO1M are shown in Fig. 10c, f, i for the years 1987, 1988 and 1989, respectively. It is seen that the simulated seasonal rainfall is more over the foothills of the Himalayas and many parts of northeast India in the case of EHO2M for all the years. It is also noticed that almost all parts of the west coast of India have received more rainfall in EHO2M experiments for all the years; however, the area with higher rainfall over the west coast is observed during the excess year. This result is in agreement with the previous studies (ABE et al.
Figure 11 Seasonal mean wind vector and magnitude (m/s) at 850 hPa a from NCMRWF model; b from NCMRWF model driven RegCM3 control run; c difference (m/s) between experiments EHO2M (10 % increase of the Himalayan height) and EHO1M (10 % decrease of the Himalayan height) calculated for the year 1987; d, e and, f are same as a, b and, c, respectively, but calculated for 1988; g, h and, i are same as a, b and, c but calculated for 1989
P. Sinha et al.
Pure Appl. Geophys.
2003; SONG et al. 2010; ZHONGFENG et al. 2010) that stronger water vapour convergence occurs due to increase of the orographical height. The seasonal mean wind vector as well as magnitude at 850 hPa obtained from the NCMRWF model, CONTM run and, wind differences between EHO2M and EHO1M for the year 1987, 1988 and, 1989 are shown in Fig. 11. It is noticed from the figure that both the models (GCM as well as RegCM) represent well the seasonal mean wind pattern at a low level associated with summer monsoon, i.e., stronger southwesterly wind associated with SJ over the Arabian Sea; however the strength of SJ is more in RegCM. Near the west coast of India, stronger southwesterly wind flows mainly through the southern part of west coast in the NCMRWF model, while
it flows through northern part of the west coast in RegCM for all three years. The wind difference between EHO2M and EHO1M (Fig. 11c, f, i) indicates that the southwesterly flow over the Arabian Sea is stronger in EHO2M than EHO1M experiments for all the years. However, the intensity in wind difference is higher during an excess year and lower during a normal year. The increased intensity of the south westerly wind because of the enhancement of the Himalayan orography is found over north of the west coast and adjoining areas of India in each individual year. It has also been noticed that the intensity of the SJ is enhanced during all the years owing to an increase in the height of the Himalayan mountain region. A stronger southwesterly due to increased orography is observed over more parts of
Figure 12 Seasonal mean zonal wind (m/s) shear from a NCEP-DOE reanalysis 2, b NCMRWF model driven RegCM3 run of EHO2M (10 % increase of the Himalayan height) and, c NCMRWF model driven RegCM3 run of EHO1M (10 % decrease of the Himalayan height) experiments calculated for the year 1987; d, e and, f are same as a, b and, c, respectively, but calculated for 1988; g, h and, i are same as a, b and, c but calculated for 1989
Role of the Himalayan Orography
Figure 13 Sectorial zonal mean velocity (m/s) from a NCEP-DOE (NNRP2) data, b NCMRWF model, c RegCM3 control run (CONTM), d EHO2M experiment and, e EHO1M experiment calculated for 1987 year; f, g, h, i and, j are same as a, b, c, d and, e, respectively but calculated for 1988 year
the west coast during an excess year than in a normal and a deficit year. As indicated in Sect. 4.2.1, the increase of the Himalayan orography strengthens the low level
circulation, and this enhancement of the intensity in low level southwesterly wind may possibly be a cause of more rainfall over the west coast of India in EHO2M than EHO1M experiments. The seasonal
P. Sinha et al.
Pure Appl. Geophys.
mean JJAS wind at 200 hPa obtained from the NCMRWF model and RegCM (Figures not shown) indicate that the intensity and location TEJ is well represented by both the models; however, the intensity is lesser in the models than the observation. It is noticed that the TEJ is stronger in the RegCM model than its driven GCM and also closer to the observations. The subtropical WJS that shifted further north during the summer monsoon is well simulated by both the GCM and RegCM for all the years; however the strength of this subtropical WJS is higher in RegCM than GCM. The wind difference at 200 hPa between EHO2M and EHO1M experiments of RegCM (Figures not shown) suggest that the tropical easterly jet becomes stronger with the enhancement of the Himalayan orography. The subtropical WJS becoming stronger near 37°N latitudinal belt with the enhancement of the orography during all the years; however, the rate of intensification in magnitude of WJS varies from year to year. It is of interest to study the zonal wind shear between 850 and 200 hPa pressure levels to understand the impact of Himalayan orography on
circulation. Seasonal mean zonal wind shear between 850 and 200 hPa pressure levels obtained from NNRP2 data and EHO2M (hereafter referred to as P10 also) and EHO1M (hereafter referred to as M10 also) of RegCM experiments for 1987, 1988 and 1989 years are shown in Fig. 12. It is seen that an area with stronger positive wind shear over the Arabian Sea, Indian region and the Bay of Bengal exists during an excess year than a normal year. The positive wind shear with lesser magnitude over a smaller area is found during a deficit year than in excess and normal years. The RegCM model in both P10 and M10 experiments are also able to represent this feature very well; however, the strength is weaker in the model than the verification when compared for the corresponding years. During all the years, the area with stronger positive wind shear is more in P10 experiments and closer to the verification analysis. Study of the vertical structure of the low level as well as upper air circulation for the two extreme rainfall years, i.e., deficit (1987) and excess (1988) has been carried out. For this purpose, a latitudinal
Figure 14 Five homogeneous rainfall regions over India based on PARTHASARATHY et al. (1995)
Role of the Himalayan Orography
cross-section of the sectorial (30°E–120°E) seasonal mean zonal wind for verification analysis (NNRP2 data), NCMRWF global model output, RegCM experiments, i.e., CONTM, EHO2M and EHO1M for 1987 and 1988 are shown in panels a–j of Fig. 13, respectively. It is seen that during the deficit year, the SJ, the TEJ and subtropical WJS in upper air are stronger during the excess year than the deficit year in the verification analysis (NCEP-DOE). As mentioned earlier, the global model has several deficiencies that get improved by the use of the RegCM. In the RegCM control run, the strength of SJ is higher than the NCMRWF model and the location of SJ is closer to verification analysis. However, stronger SJ during excess year is not represented in the CONTM run. The core wind speed of the subtropical WJS is higher in a CONTM run than the NCMRWF model and closer with the verification analysis for an excess year. However, during a deficit year, the core wind speed of subtropical WJS is overestimated by the RegCM model (CONTM). It is seen from the figure that the area with stronger subtropical WJS is more during an excess year than a deficit year in CONTM. The strength of the TEJ is well simulated by CONTM in an excess year, while it is overestimated in a deficit year. During an excess year, EHO2M shows a larger area with stronger SJ than in a deficit year and this feature is not represented by EHO1M. It is seen that the area with stronger SJ from EHO2M is closer to verification analysis than from EHO1M. However, the strengths of the TEJ and subtropical WJS obtained from EHO2M and EHO1M experiments are almost similar with the CONTM experiment. In order to evaluate RegCM model performance in simulating precipitation during summer monsoon season, a detailed examination of rainfall distribution and intensity obtained from various experiments of RegCM over India as well as five homogeneous regions has been carried out. Based on PARTHASARATHY et al. (1995), the five homogeneous regions namely northwest India (NWI), west central India (WCI), southern peninsular India (SPI), central north east India (CNEI) and, north east India (NEI) are considered and shown in Fig. 14. The RegCM model simulated precipitation is interpolated linearly to IMD grid points before carrying out the analyses.
Figure 15 Area average of seasonal rainfall (mm/day) calculated over India as well as five rainfall homogeneous regions namely northwest India (NWI), west central India (WCI), southern peninsular India (SPI), central north east India (CNEI) and, north east India (NEI) obtained from IMD, NCMRWF driven RegCM simulations with control height (Cont), 10 % increase (EHO2M experiments; P10) and 10 % decrease (EHO1M experiments; M10) of the Himalayan orography computed for a 1987 year, b 1988 year and, c 1989 year
Area averaged seasonal rainfall over all-India as well as five homogeneous rainfall regions obtained from IMD and RegCM simulated results for control, EHO2M (P10) and EHO1M (M10) computed for three different years is shown in Fig. 15. It is seen that area averaged seasonal rainfall over all-India is underestimated in all the experiments of RegCM simulations, however results from P10 is closer with the observations during all the years. It is found that
P. Sinha et al.
Pure Appl. Geophys.
Table 2 Statistical relationship [spatial correlation coefficient (CC), Root mean square error (RMSE) and mean bias (Bias)] between observed (IMD) rainfall and model results [GCM, P10 (EHO2M) and, M10 (EHO1M) experiments] calculated over India and five homogeneous regions for 1987, 1988 and 1989 years Correlation coefficient
t99
1987
INDIA NWI WCI SPI CNEI NEI
1988
1989
Control
P10
M10
Control
P10
M10
Control
P10
M10
0.38 -0.33 0.49 0.71 0.37 0.44
0.4 -0.27 0.45 0.82 0.28 0.44
0.34 -0.44 0.41 0.75 0.34 0.5
0.4 0.25 0.48 0.84 0.01 0.24
0.48 0.38 0.56 0.9 0.06 0.24
0.44 0.1 0.49 0.8 0.01 0.29
0.29 0.16 0.26 0.83 0.3 0.24
0.25 0.05 0.28 0.81 0.31 0.21
0.22 0.11 0.28 0.64 0.3 0.28
0.05 0.19 0.22 0.26 0.19 0.24
Root mean square error (RMSE) 1987
INDIA NWI WCI SPI CNEI NEI
1988
1989
Control
P10
M10
Control
P10
M10
Control
P10
M10
6.77 3.13 5.29 4.9 6.38 12.27
6.75 3.33 6.24 4.21 6.5 11.63
7.02 3.16 5.61 4.61 6.57 13.46
7.33 3.47 7.49 7.87 6.69 12.36
6.82 3.1 7.23 5.97 6.9 11.32
7.31 3.69 7.79 7.58 6.73 12.16
6.43 2.27 7.39 7.53 7.36 9.03
6.78 2.37 7.89 7.9 7.4 9.37
6.56 2.36 7.41 8.33 7.12 9.39
Bias (%) 1987
INDIA NWI WCI SPI CNEI NEI
1988
1989
Control
P10
M10
Control
P10
M10
Control
P10
M10
-37.72 22.19 -26.19 -30.56 -68.89 -61.48
-30.1 36.68 -10.14 -30.57 -68.28 -53.03
-44.63 22.56 -32.69 -29.91 -72.69 -71.04
-55.31 -38.19 -45.27 -56.15 -67.54 -67.36
-46.33 -26.98 -37.61 -43.21 -67.54 -54.37
-57.18 -45.52 -50.74 -51.99 -70.26 -67.02
-50.11 -12.36 -53.55 -59.02 -76.8 -52.5
-44.9 -14.9 -40.86 -63.94 -77.02 -46.49
-52.01 -10.47 -57.72 -65.5 -74.24 -59.62
around a 17 % increase of seasonal rainfall over the Indian domain occurred with the enhancement of the Himalayan orography by 10 % from its control height. The increase in simulated rainfall in P10 than in control experiments is also noticed almost over all homogeneous regions during all the years except over the region CNEI. However, the rate of increase is higher in WCI (*20 %) and NEI (*28 %) regions than other parts of India. Several statistical analyses such as spatial correlation coefficient (CC) and root mean square error (RMSE) and bias (%) of seasonal rainfall are carried out between the RegCM model output and IMD
observation for three different years 1987, 1988 and 1989 and are provided in Table 2. Statistically significant values at 99 % confidence level (CL) for India and five homogeneous regions are also given in the last column of Table 2. The spatial CC obtained from all experiments is statistically significant at a 1 % significant level (SL) over India for all the years. Statistically significant (at 1 % SL) CC is noticed for all the experiments in all years over WCI and SPI homogeneous regions. The spatial CC is higher in P10 experiments than control and M10 experiments over India in 1987 and 1988 years. However, CC is less in P10 experiments than the control in the
Role of the Himalayan Orography
1989 year. An increase of CC is noticed for NWI, SPI regions in 1987 and 1988 years for P10 experiments. The RMSE is reduced in P10 experiments during 1987 and 1988 years over India, SPI and NEI regions. However, RMSE is higher in the P10 experiment over the CNEI region during all the years. It is noticed from the Table 2 that the bias (in %) is reduced notably in P10 experiments during all the years over all-India as well as WCI and NEI homogeneous regions. However, no notable improvement is observed in P10 experiments over the CNEI region. The poor performance over the CNEI region is probably due to a poor representation of the monsoon trough as well as the omega pattern in the upper air by the RegCM model.
5. Conclusions The regional climate model RegCM has been utilized to examine the sensitivity of the Himalayan orography representation in the simulation of circulation and rainfall for three monsoon seasons. Three distinct monsoon seasons such as deficit (1987), excess (1988) and, normal (1989) rainfall monsoon years are considered for this study. The RegCM is forced by (1) observed reanalysis data (NNRP2) and (2) NCMRWF model output. The performance of the RegCM model driven by observed analysis, i.e., NNRP2 has reasonable skill in simulating mean JJAS circulations over the monsoon regions and associated rainfall distribution patterns over India. RegCM has the capability to simulate semi-permanent features such as low level circulations including a stronger wind at the Somali coast and upper air circulations associated with tropical easterly jet, and a subtropical westerly jet stream associated with the Indian summer monsoon during all the years. The performance of RegCM is adequate in representing the variation in circulation and precipitation pattern and intensity during excess and deficit years. The performance of the RegCM model driven by the NCMRWF model is satisfactory in simulating mean JJAS rainfall distribution patterns over India during all the years. The skill of the RegCM is better than the global model in simulating the rainfall
distribution and intensity over some parts of India such as southern parts of the west coast, many parts of north India especially over north of the Gangetic plain and northeast India. With this, the representations of some summer monsoon semi-permanent features are better in the RegCM model than the global model. However, the RegCM model at coarse resolution has a weakness in representing the monsoon trough as well as the vertical pressure velocity pattern over central and east India. The RegCM model driven by the NCMRWF global model has the potential in simulating the excess and deficit years and representing the variation. It is noted here that the representation of the Himalayan orography is less in the RegCM model (control) than the actual height of the Himalayas. Thus, the influence of the Himalayan orography has been studied by decreasing and increasing of height by 10 % from the mean height of RegCM. The driving force to the regional model is provided from NNRP2 as well as NCMRWF GCM. The simulated rainfall over the whole ofIndia especially, the foothill of the Himalaya, west coast of India and over the north east India is more when the height of orography is enhanced. This is may be because of stronger water vapour convergence due to increase of the orographical height that enhances the local convective activities. Owing to an increase of the Himalayan orography, the divergence of upper air uplifted and the air has more potential energy; when this upper air sinks over the ocean to complete its cycle, the potential energy converted into kinetic energy that increases the strength of the low level circulation. Thus, the low level southwesterly wind including SJ stream becomes stronger with the enhancement of the Himalayan orography. As a result of this, the rainfall is more over the west coast of India. The upper air circulation associated with the tropical easterly jet stream is stronger in case of enhanced orography of the Himalaya mountain region. These results are supported by the previous studies (ROY ABRAHAM et al. 1996; ABE et al. 2003; MOHANDAS et al. 2009; SONG et al. 2010; ZHONGFENG et al. 2010) that the large scale fields such as low level wind including SJ and upper air circulations related to the tropical easterly jet stream associated with the Indian summer monsoon becomes stronger with the increased
P. Sinha et al.
Pure Appl. Geophys.
orography. The enhancement in the precipitation amount is high for steeper orography representation during the stronger monsoon season. This is probably due to the strength of the low level south westerly wind is more during an excess year. During an excess year, the impact of steeper orography is more and as a result, simulated precipitation is higher near the foothills of the Himalayas. Several statistical analyses reveal that the increase in the height of the Himalayan orography by 10 % from mean height in RegCM is able to represent better monsoon rainfall distribution and intensity over India as well as many homogeneous regions. Therefore, the representation of the Himalayan orography in regional climate model is one important issue in the simulation of the Indian summer monsoon.
regional climate model: Domain size experiments. Clim. Dyn., 12, 573–587. BOOS, W.R., and Z. KUANG, 2010: Dominant control of the South Asian monsoon by orographic insulation versus plateau heating. Nature, 463, 218–222. doi:10.1038/nature08707. CHAKRABORTY, A., R.S. NANJUNDIAH, and J. SRINIVASAN, 2002: Role of Asian and African orography in Indian summer monsoon. Geophys. Res. Lett., 29, doi:10.1029/2002GL015522. COOK, K.H., and I.M. HELD, 1992: The stationary response to large scale orography in a general circulation model and a liner model. J. Atmos. Sci., 49, 525–539. DAS, P.K., and H.S. BEDI, 1978: The inclusion of Himalayas in a primitive equation model. Indian J. Meteorol. Hydrol. Geophys. (Currently: Mausam), 29, 373–383. DICKINSON, R. E., 1995: Land-atmosphere interaction. Rev. Geophys., 33, 917–922. FENNESSY, M.J., J.L. KINTER, B. KIRTMAN, L. MARX, S. NIGAM, E. SCHNEIDER, J. SHUKLA, D. STRAUS, A. VERNEKAR, Y. XUE, and J. ZHOU, 1994: The Simulated Indian Monsoon: A GCM Sensitivity Study. J. Clim., 7, 33–43. GIORGI, F., and L.O. MEARNS, 1991: Approaches to the simulation of regional climate change: A review. Rev. Geophys., 29, 191–216. GIORGI, F., and R. AVISSAR, 1997: The representation of heterogeneity effects in earth system modeling: Experience from land surface modeling. Rev. Geophys., 35, 413–438. GRELL, G.A., J. DUDHIA, and D.R. STAUFFER, 1994: Description of the fifth generation Penn State/NCAR Mesoscale Model (MM5). Tech. Rep. TN-398 ? STR, NCAR, Boulder, Colorado, pp. 1–12. GROSSMAN, R.L., and D.R. DURRAN, 1984: Interaction of low level flow with the Western Ghat mountains and off shore convection in the summer monsoon, Mon. Wea. Rev., 112, 652–671. HAHN, D.G., and S. MANABE, 1975: The Role of Mountains in the South Asian Monsoon Circulation. J. Atmos. Sci., 32, 1515–1541. JI, Y., and A.D. VERNEKAR, 1997: Simulation of the Asian summer monsoons of 1987 and 1988 with a regional model nested in a global GCM. J. Clim., 10, 1965–1979. KANAMITSU, M., W. EBISUZAKI, J. WOOLLEN, S.-Y. YANG, J.J. HNILO, M. FIORINO, and G.L. POTTER, 2002: NCEP-DEO AMIP-II Reanalysis (R-2). Bull. Am. Meteorol. Soc., 83, 1631–1643. KAR, S. C., 2007: Global model simulations of interannual variability of the Indian summer monsoon using observed SST variability. NCMRWF Research Report, NMRF/RR/2/2007; 40 pp. KAR, S.C., G.R. IYENGAR, and A.K. BOHRA, 2011: Ensemble spread and model systematic errors in the monsoon rainfall forecasts using the NCMRWF global ensemble prediction system, Atmosfera, 24, 173–191. KASAHARA, A., 1980: Influence of orography on the atmospheric general circulation. WMO Report, Orographic Effects in Planetary Flows, pp 1–49 (SEE N80-33811 24–42). KRISHNAMURTI, T.N., K. INGELS, S. COCKE, T. KITADE, and R. PASCH, 1984: Details of low latitude medium range numerical weather prediction using a global spectral model, Part II: Effects of orography and physical initialization. J. Meteorol. Soc. Jpn., 62, 613–648. MOHANDAS, S., S.K. DASH, and P.K. MOHANTY, 2009: Indian summer monsoon simulation studies with different orographic representations in a spectral GCM. Int. J. Climatol., 29, 269–288. PAL, J.S., F. GIORGI, X. BI, N. ELGUINDI, F. SOLMON, X. GAO, S. A. RAUSCHER, R. FRANCISCO, A. ZAKEY, J. WINTER, M. ASHFAQ, F.
Acknowledgments This research work is financially supported from the Department of Agriculture and Cooperation (DAC), Ministry of Agriculture, Govt. of India through the project entitled ‘‘Development and Application of Extended Range Forecast System for Climate Risk Management in Agriculture’’ at IIT, Delhi. The authors also acknowledge with thanks to IMD for providing gridded (1° 9 1°) rainfall data. The authors also duly acknowledge NCEP for reanalysis 2 data and the NOAA for optimum interpolated SST version 2 data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at http://www.esrl.noaa.gov/psd/. REFERENCES ABE, M., A. KITOH, and T. YASUNARI, 2003: An Evolution of the Asian Summer Monsoon Associated with Mountain Uplift— Simulation with the MRI Atmosphere-Ocean Coupled GCM. J. Meteorol. Soc. Japan., 81, 909–933. AN, Z.S., J. E. KUTZBACH, W.L. PRELL, and S.C. PORTER, 2001: Evolution of Asian monsoons and phased uplift of the HimalayaTibetan Plateau since later Miocene times. Nature, 411, 62–66. doi:10.1038/35075035. ANTHES, R. A., 1977: A cumulus parameterization scheme utilizing a one-dimensional cloud model. Mon. Wea. Rev., 105, 270–286. BANERJEE, S. K., 1929: The effect of the Indian mountain ranges on the configuration of the isobars. Indian J. Phys., IV, 477–502. BHASKARAN, B., R.G. JONES, J.M. MURPHY, and M. NOGUER, 1996: Simulations of the Indian summer monsoon using a nested
Role of the Himalayan Orography S. SYED, J.L. BELL, N.S. DIFFENBAUGH, J. KARMACHARYA, A. KONARE, D. MARTINEZ, R.P.D. ROCHA, L.C. SLOAN, and A. L. STEINER, 2007: Regional climate modelling for the developing world: The ICTP RegCM3 and RegCNET. Bull. Am. Meteorol. Soc., 88, 1395–1409. PARTHASARATHY, B., A.A. MUNOT, and D.R. KOTHAWALE, 1995: Monthly and Seasonal Rainfall Series for All-India Homogeneous Regions and Meteorological Subdivisions: 1871–1994. Contributions from Indian Institute of Tropical Meteorology. Research Report RR-065 (ISSN 0252-1075), Aug. 1995, Pune 411 008 INDIA. PARTHASARATHY, B., A.A. MUNOT, and D.R. KOTHAWALE, 1994: All India monthly and seasonal rainfall series 1871–1993. Theor. Appl. Climatol., 49, 217–224. PIELKE, R., and R. AVISSAR, 1990: Influence of landscape structure on local and regional climate. Landscape Ecol., 4, 133–155. PIELKE, R., 2001: Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. Rev. Geophys., 39, 151–177. PRELL, W.L., and J.E. KUTZBACH, 1992: Sensitivity of the Indian monsoon to forcing parameters and implications for its evolution. Nature, 360, 647–652. RAJEEVAN, M., J. BHATE, J. KALE, J., and B. LAL, 2006: High resolution daily gridded rainfall data for the Indian region: analysis of break and active monsoon spells. Curr. Sci., 91, 296–306.
ROY ABRAHAM, K., S.K. DASH, and U.C. MOHANTY, 1996: Simulation of Monsoon circulation and cyclones with different types of orography. Mausam, 47, 237–250. SINGH, H.N., S.D. PATIL, S.D. BANSOD, and N. SINGH, 2011: Seasonal variability in mean sea level pressure extremes over the Indian region. Atmospheric Res, 101, 102–111. SINHA, P., U.C. MOHANTY, S.C. KAR, S.K. DASH, and S. KUMARI, 2012a: Sensitivity of the GCM driven summer monsoon simulations to cumulus parameterization schemes in nested RegCM3. Theor. Appl. Climatol., doi:10.1007/s00704-012-0728-5. SINHA, P., U.C. MOHANTY, S.C. KAR, S.K. DASH, A. ROBERTSON, and M. TIPPETT, 2012b: Seasonal Prediction of the Indian Summer Monsoon Rainfall using Canonical Correlation Analysis of the NCMRWF Global Model Products. Int. J. Climatol., doi: 10.1002/joc.3536. SONG, J.-H., H.-S. KANG, Y.-H. BYUN, and S.-Y. HONG, 2010: Effects of the Tibetan Plateau on the Asian summer monsoon: a numerical case study using a regional climate model. Int. J. Climatol., 30, 743–759. ZHONGFENG, X.U., Q. YONGFU, and F.U. CONGBIN, 2010: The Role of Land–sea Distribution and Orography in the Asian Monsoon. Part II: Orography. Adv. Atmos. Sci., 27, 528–542.
(Received September 12, 2012, revised April 11, 2013, accepted April 18, 2013)