Nmm project Proposal Title of the proposed project



Download 303.71 Kb.
Page1/6
Date18.10.2016
Size303.71 Kb.
  1   2   3   4   5   6


NMM Project Proposal


  1. Title of the proposed project:

Towards understanding the biases in the model SST, wind field and rainfall in the Climate Forecasting System for the Monsoon - Indian Ocean domain (30E, 120E, 30S & 30N).


  1. Brief information about Principal Investigator (PI) and Co-PI(s) :

PI :

Name : Dr.SSVS Ramakrishna

Date of birth : 18-9-1959

Institution: Department of Meteorology and Oceanography Andhra University

Qualification : M.Sc, Ph.D
Co- PI :

Name: Dr. C.V.Naidu

Date of birth: 1-6-1964

Institution: Department of Meteorology and Oceanography, Andhra University

Qualification: M.Sc, Ph.D.
Consultant 1:

Name : Dr. R.R.Rao

Date of birth : 16-4-1949

Institution : Department of Meteorology and Oceanography, Andhra University

Qualification : M.Sc(Tech.), Ph.D.
Consultant 2:

Name : Dr V. B. Rao

Date of birth : 16-10-1941

Institution : Department of Meteorology and Oceanography, Andhra University

Qualification : M.Sc(Tech.), Ph.D.
Consultant 3:

Name : Dr. D.V. Bhaskar Rao

Date of birth : 30-9-1948

Institution: Department of Meteorology and Oceanography, Andhra University

Qualification: M.Sc (Tech.), Ph.D.


  1. Project Summary :

1. Proposed work

1.1 Intellectual merit of the proposed work

The Climate Forecast System (CFS) developed at NCEP is a fully coupled ocean-land-atmosphere model and presently implemented and used for making long-term integrations at IITM for the assessment of dynamical seasonal prediction of the Indian Summer Monsoon (ISM) as part of the Indian Monsoon Mission (IMM) . The NCEP CFS model is being run at T382 (~100 km) horizontal resolution coupled with MOM4 ocean model and dynamical sea ice and land surface models. A suite of model integrations have been completed with different land surface processes, cumulus convection schemes, PBL schemes, and sea-ice model that yielded daily outputs as related to the different experiments of the physical processes. It is important to assess the performance of the model in forecasting rainfall on different time scales for different forecast lead times. The model outputs for retrospective forecasts for the recent 28 years that are available at IITM show large biases in the wind field, rainfall and SST. So there is a strong need to characterize and understand the causative factors that contribute to these biases in the wind field, rainfall and SST on different time scales for different lead times of integration. These model outputs and the corresponding observed fields would be analyzed to assess the model performance in the prediction of ISM on different time scales.
1.2 Broader Impact of the proposed work
The evolution of ISM is being predicted on hindcast mode with initial conditions taken from the preceding January, February, March, April and May conditions. Detailed analysis of archived model outputs obtained for different numerical experiments would reveal the performance of the model in reproducing the observed variability in the wind field, rainfall and SST. This study would lead to identify the characteristics and possible causes for the bias and suggest remedial measures for probable improvement of the prediction of ISM rainfall. These biases in the fields such as winds, SST and rainfall will throw light on the limitations of the model modules which do not perform satisfactorily. This study would be helpful for the model improvement.
Project Description:

2. Research Objectives

The project envisages to address the following :



  • Characterization of biases in the wind field, SST and rainfall and their spatio-temporal variability

  • Identification of causative mechanisms for the observed biases in the wind field, SST and rainfall

  • Suggestion of remedial measures to reduce these biases


3. Statement of Work (methodology to be adopted)
Module 1: Towards understanding the bias in SST in the Climate Forecasting System model (Consultant: Dr R R Rao)
a) Possible reasons for the observed bias in SST simulation
Many atmosphere-ocean coupled models simulate either warm or cold SSTs in the tropical ocean basins. The Arabian Sea SST biases are common in coupled models and may therefore influence the monsoon and its sensitivity to climate change. The probable causative mechanisms that contribute to the observed biases in SST are attributed to inaccurate parameterization of the atmospheric forcing fields and the oceanic processes. Some processes such as solar radiation and advection of warm waters contribute to heating while net long wave radiation, evaporation, sensible heat flux, advection of cooler waters, entrainment and upwelling contribute to cooling and hence these processes must be quantified accurately. In regions of large fresh water forcing - occurrence of barrier layer must be simulated accurately. Therefore in the Bay of Bengal and the eastern equatorial Indian Ocean - the salinity effects on SST need to be simulated properly. The SST is also sensitive to accurate definition of MLD. The evolution of MLD is controlled by both surface meteorological forcing and as well as by the evolution of pycnocline influenced by surface wind stress curl and propagating long period waves and this needs to be simulated accurately. Therefore errors in the above mentioned fields and physical processes will lead to biases in SST.
b) Present proposal
The IITM has carried out simulations with the CFS atmosphere-ocean coupled model for retrospective forecasts (recent 28 years) for different lead times. The model outputs show biases in SST with variable signs and magnitudes across the basin. In this study, it is proposed to address the following issues for the TIO:


  • Characterization of SST bias and its spatio-temporal variability

  • Identification of causative mechanisms for the observed SST bias

  • Suggestion of remedial measures to reduce the observed SST bias

This proposal aims to address the spatial structure of the climatology, intraseasonal and interannual variability of SST bias through comparison with observations/analysis and identify possible mechanisms that contribute to the observed bias in SST through diagnostic studies of the model forcing fields and model outputs generated under different experiments. All the available in situ and satellite data products on surface winds, fluxes, subsurface temperature, salinity and current structures would be utilized to compare and evaluate the performance of the coupled model using approaches such as pattern correlation, RMS error/bias and absolute error. The local and remote mechanisms governing low frequency SST variability will be examined in detail.
c) Expected Outcome

The results from the diagnostic studies of the model forcing fields and model outputs would indicate the model deficiencies in the simulation of the SST on different time scales. The space-time variability of SST bias across the basin for different lead times would be known. Analysis of the surface forcing fields and model simulated ocean outputs would reveal their corresponding deficiencies. Then necessary improvements that are required to redefine the surface forcing fields, parameterization of various governing mechanisms towards more accurate simulation of SST in the TIO would be identified. Improvements in physics and coupling would lead to improved simulation of SST. This would be a challenging task given the complexity of dynamical and thermodynamical coupled physical processes in the TIO. Identification and implementation of corrective measures to improve the SST biases should be a clear target for the model development.


Module 2: Towards understanding the bias in the wind field in the Climate Forecasting System model (Consultant: Dr V B Rao)
a) Possible reasons for the observed bias in wind field simulation
The wind field is an important component of the monsoons. The monsoon weather systems are characterized with the knowledge of 3-D distribution of winds. Rapid changes occur in the wind field and the associated vorticity fields in association with the onset and progress (active-break cycles, monsoon lows and depressions) of the monsoons. Hence accurate simulation of wind field assumes special significance. The low level winds are of vital importance for several monsoon dynamical processes. The upper level wind fields over the Indian subcontinent also have a strong bearing on several aspects of the monsoon related phenomena. During the onset regime of the monsoon, the rapid establishment of a steady westerly winds over the monsoon domain coincides with the abrupt beginning of rainy season over the southern tip of the Indian peninsula. The wind field in the lower troposphere is controlled by processes in the PBL and cumulus convection in the atmosphere. Accurate simulation of the 3-D distribution of temperature and moisture in the troposphere would set the stage for accurate simulation of wind field. This indicates that in the tropics accurate simulation of surface pressure fields and rainfall (and hence SST) will result in the simulation of accurate wind fields.
b) Present Proposal:

We propose to characterize the biases in the modeled wind field with regard to seasonal mean, intraseasonal variability and interannual variability. The modeled wind field will be evaluated for different phases of the monsoon such as onset, active-break cycles, monsoon lows and depressions and monsoon withdrawl with the available observations/analysis. We will also look at the impact of interannual signals such as ENSO and IOD on the general behavior of the monsoon both in model output and observations. We will also attempt to identify possible reasons for the disagreement. Low level wind at 850 hPa variations are strongly related to the pressure gradient changes, which are in turn related to the low level temperature variations as such as changes in SST. Thus we try to relate biases determined in SST and wind. The low level winds are also strongly related to the moisture transport from both the Arabian Sea and Bay of Bengal, which are in turn vital for the rainfall over the continent. We propose to determine the biases in the wind determined by the CFS system, both direction and speed. We propose to relate these biases to the boundary layer and moist atmospheric convection schemes used in the CFS model. We also try to make an attempt to determine the sensitivity of these biases to the choice of PBL and cumulus convection schemes. To determine these biases quantitatively we use the standard statistical methods such as the RMSE, absolute error, pattern correlation and the bias.


c) Expected outcome:

The biases in the CFS model winds will be characterized. The causative mechanisms that contribute to these biases will be identified. The possible reasons for the biases in the upper tropospheric winds in particular the TEJ strength will be determined through simple relations such as thermal winds associated with the temperature gradients. Suggestions will be made to correct these biases which are expected to improve the model winds. It is proposed to examine any trends in the CFS output. An attempt will also be made to rectify these biases through PBL and cumulus convection schemes.


Module 3: Towards understanding the bias in rainfall in the Climate Forecasting System model (Consultant: Dr D V B Rao)
a) Possible reasons for the observed bias in rainfall simulation
The GCMs are constructed to simulate the environmental conditions such as horizontal and vertical distributions of wind, temperature and humidity, and SST through interactive feedback mechanisms. The reasons for deficiencies in simulating rainfall could be due to systematic errors in large scale circulation to which convection responds and/or due to incorrect responses of the model convection scheme to large-scale environment. Between these two scenarios, the reasons could be more complex because of feedbacks between the convection and environment and between different physical parameterizations. Our aim is to identify whether biases arise from an unrealistically simulated large-scale environment, from parameterization errors, or from some complex interaction between the two. If the errors are due to parameterization we aim to investigate how the parameterizations might be improved.
In this research part, we will examine the biases in the annual cycle of precipitation from the several experiments with CFS model to diagnose the relationship between precipitation and several environmental variables, identify common behaviors with common biases in precipitation and thus identify to what atmospheric properties precipitation is most sensitive in the CFS model. This will be achieved through preparing monthly mean climatology and comparing with NCEP reanalysis for environmental fields and gridded rainfall for precipitation. The formation and movement of the monsoon trough during the onset phase and the intraseasonal oscillations of the monsoon trough during the months of July and August will be examined for identification of bias in environmental conditions. Correspondingly rainfall analysis will be made to correlate the rainfall zones and convergence regions.
b) Present proposal
The characteristics of the model derived daily precipitation during ISM season (i.e.) the four months of June-July-August-September over the different sub-regions of India from the NCEP CFS would be studied. The model derived daily rainfall will be compared to gridded rainfall data. Differences in the occurrence of daily rainfall between the observations and corresponding CFS forecasts will be examined as a function of forecast lead-time for the 28-year model run period. The bias in CFS will be quantified by comparing the observations to the CFS forecasts on grids. Enough care will be taken to see that the quantitative nature of the bias in CFS is preserved although some spatial details of the gridded observations could be lost.
An attempt would be made to understand how the different experiments reproduce the variability of daily rainfall. Model daily rainfall will be compared with IMD and APHRODITE gridded rainfall. We will examine if there is a consistent relationship between biases in the distribution of daily rainfall and biases in the monthly mean, for eg. to see if the model underestimates the frequency of heavy-rain events and also the monthly accumulation.
The temporal characteristics of daily rainfall will be diagnosed through autocorrelation of daily precipitation computed for each grid point and then averaged over the grid points. We will analyze the model simulated large-scale environment from the model experiments and the NCEP reanalysis to see biases in the large-scale environment that are consistent with the biases in precipitation. For example, conservation of moisture will prescribe that the column-integrated moisture convergence will be equal to precipitation minus evaporation (P – E) and condensational heating exceeding radiative cooling will be associated with large-scale ascent; the stronger the precipitation, the stronger the ascent. This will provide an understanding of the predictability of surface mass convergence and precipitation. The simulated relationship between surface convergence and precipitation is important in determining what factors control precipitation in the models. Current theories of tropical precipitation can be organized in two ways, one is that the convergence of the low level winds determine the location and intensity of precipitation and the influence of SST is through its control on the winds via the momentum budget and the other is that precipitation is determined locally by thermodynamic factors such as boundary layer entropy or moist static energy.
Apart from the relationship between precipitation and large scale environment, local relationships may also exist (i.e.) the precipitation at any given point could be related the vertical profiles of temperature and humidity at that location. We will analyze the vertical profiles of temperature and humidity in dry and rainy conditions as we expect that the difference in vertical profiles for rainy and dry days reflects climatological differences, more than day-to-day variability.
The frequency of wet and dry spells (consecutive wet and dry days) will be examined. Similarly daily rainfall statistics in relation to ENSO phases (cold and warm phases) will also be studied. A classification of historical warm (El Niño) and cold (La Niña) episodes developed by the National Oceanic and Atmospheric Administration’s (NOAA) Climate Prediction Center will be used to develop composites of the observed daily rainfall statistics by ENSO phase. El Niño and La Niña episodes are identified using the Oceanic Niño Index (ONI; Kousky and Higgins 2007), which is one of the principal measures used by NOAA for monitoring and assessing ENSO. The NOAA operational definitions of El Niño and La Niña are keyed to the ONI (Kousky and Higgins 2007): El Niño is characterized by a positive ONI ≥ +0.5°C; La Niña is characterized by a negative ONI ≤ -0.5°C. These definitions will help to properly identify all historical warm and cold episodes. For comparison purposes, a similar procedure is used to identify ENSO events in the CFS coupled simulations. Events are chosen using a threshold of ±0.5°C for 3-month running mean SST anomalies in the Niño-3.4 region (5°N–5°S, 120°–170°W). The model bias in the estimation of precipitation will be studied with respect to different sub-regions. This type of study will bring out the seasonal dependence of the bias in the CFS forecasts. This study also aims at assessing the ability of the CFS to produce the observed daily precipitation characteristics over different sub-regions.
c) Expected Outcome
The bias in rainfall prediction on different time scales will be obtained as per the methodology presented in the preceding section. For this purpose model outputs derived from experiments with different schemes of land surface processes, cumulus convection, PBL and Sea-Ice schemes as archived at IITM will be used along with NCEP reanalysis fields and IMD/APHRODITE gridded rainfall data over land and satellite derived rainfall products (TRMM/ CMAP/ GPCP) over ocean region. The analyses is expected to provide an understanding of the reasons for the biases towards improvement of some of the dynamical and/or physical processes of the CFS model all to improve the prediction of ISMR.
Schedule (Year wise)

Year

Activity

Expected Outcome

Year - 1

  • Appointment of project personnel

  • Acquisition of CFS model outputs

  • Acquisition of observed in situ and satellite data sets and analysis fields

  • Acquisition of computer hardware and software

  • Development of computer codes and graphical packages to characterize errors/biases

  • Assembly of model outputs and data sets

  • Sample codes for analysis of errors/biases

Year - 2

  • Analysis of model outputs to characterize biases in wind field, SST and rainfall

  • Diagnostic studies to identify possible reasons for the observed biases in wind field, SST and rainfall


  • Sample codes for analysis of errors/biases

  • Understanding of deficiencies in boundary forcings/ physical processes that lead to observed biases

Year - 3

  • Completion of tasks shown for Year -2

  • Identification of remedial measures to reduce biases

  • Documentation of results for project report and journal publications

  • Knowledge of biases in wind field, SST and rainfall

  • Improvements to reduce the biases wind field, SST and rainfall





1.1 Team Composition and expertise

Investigator

Qualifications

Field of Expertise

PI: Dr SSVS Ramakrishna

M.Sc PhD

Numerical modeling of tropical cyclones and monsoon energetics

Co-PI : Dr CV Naidu

M.Sc PhD

Monsoon diagnostics, interannual variability of Indian monsoon rainfall

Consultant 1: Dr. R R Rao

M.Sc (Tech) PhD

Indian Ocean warm pool, mixed layer dynamics

Consultant 2: Dr V B Rao

M.Sc (Tech) PhD

Tropical meteorology and climate dynamics

Consultant 3: Dr D.V.Bhaskar Rao

M.Sc (Tech) PhD

Numerical modeling of tropical cyclones, general circulation modeling

Collaborator : Prof Tetsuya Takemi, Kyoto University, Japan

Ph D

Tropical meteorology and modeling


Connections to Operational Forecast and Human Resource Development

A good outcome of the seasonal forecast will always help the Indian economy. Also a thorough understanding of the intraseasonal, interannual and spatial variability will help for the better estimates of the rainfall. Development of a thoroughly tested coupled model with improved hindcast skill suitable for the tropical Indian subcontinent is always desirable for use by the national meteorological agency i.e., India Meteorology Department. This project serves as a capacity building platform to train research associates and research fellows.


4. Results from prior MoES support (if any) : NO

[Describe any prior MoES funded work by the PI, Co-PI(s)]




Investigator

MoES grant no.

Title

Year

Description

PI













Co-PI














Facilities available at the workspace One IBM Work Station and 2 PCs


  1. Budget requirements (with justifications)

  1. Emoluments for research personnel, technical and administrative support




    1. Budget requirement for Key personnel




BUDGET




Designation

Monthly Emoluments

1st year


2nd year

3rd year

Total

Consultant 1

25000

300000 (12)

300000 (12)

3,60,000(12)

9,60,000(36)

Consultant 2

25000

300000 (12)

300000 (12)

3,60,000(12)

9,60,000(36)

Consultant 3

25000

300000 (12)

300000 (12)

3,60,000(12)

9,60,000(36)

Total

25000

9,00,000

9,00,000

10,80,000

28,80,000




    1. Budget requirement for other personnel :




BUDGET




Designation

Monthly Emoluments

1st year


2nd year

3rd year

Total

RA

22000

2,20,000 (12)

2,20,000 (12)

2,20,000 (12)

7,92,000

3 JRF’s

I&II years JRF (18,400)-III year SRF(20,700)

6,62,400(12)

6,62,400(12)

7,45,200(12)

20,70,000

Scientist/

Engineer



20,000

2,40,000(12)

2,40,000(12)

2,40,000(12)

7,20,000

Total




11,02,400

11,02,400

1118400

35,82,000




  1. Budget requirement for Travel

    1. Budget for Travel :




BUDGET

(In rupees)

1st year

2nd year

3rd year

Total

1.

Travel

1,00,000

1,00,000

1,00,000

3,00,000




Total

1,00,000

1,00,000

1,00,000

3,00,000


Download 303.71 Kb.

Share with your friends:
  1   2   3   4   5   6




The database is protected by copyright ©ininet.org 2020
send message

    Main page