Government of Hydrology Project II Water Resources Department IBRD Loan No: 4749-IN

Real Time Streamflow Forecasting and Reservoir Operation System for Krishna and Bhima River Basins in Maharashtra (RTSF & ROS)

Interim Report March 2012

DHI Water Environment Health

Real Time Streamflow Forecasting and DHI () Water & Reservoir Operation System for Krishna and Environment Pvt Ltd Bhima River Basins in Maharashtra (RTSF & 3rd Floor, NSIC Bhawan, Okhla ROS) Industrial Estate New Delhi 11 00 20

India

Tel:+911147034500 +91 11 4703 4500 Interim Report Fax:+911147034501 +91 11 4703 4501 [email protected] March 2012 www.dhigroup.com Client Client’s representative

Chief Engineer, Planning & Hydrology Superintending Engineer

Project Project No Real Time Streamflow Forecasting and Reservoir Operation System for Krishna and Bhima River Basins in 63800247 Maharashtra (RTSF & ROS)

Authors Date: Guna Paudyal 27 March 2012 Finn Hansen Gregers Jorgensen Approved by Dhananjay Pandit Hans G. Enggrob

Revision Description By Checked Approved Date Key words Classification Real Time, Streamflow, Flood, Forecasting, Open Reservoir Operation, Forecast Models, Hydrology, Hydraulics, River Basin, Capacity Building Internal

Proprietary

Distribution No of copies Client: PDF File 10 DHI: PDF file (+1 CD)

Krishna & Bhima River Basins RTSF & ROS

List of Acronyms and Abbreviations

BSD Basin Simulation Division CWC Central Water Commission DA Data Assimilation DAS Data Acquisition System DEM Digital Elevation Model DSS Decision Support System FCL Flood Control Level FCS Full Climate Station FMO Flood meteorological Office (of IMD) GIS Geographic Information System GMRBA Godavari Marathwada River Basin Agency GMS Geostationary Meteorological Satellite GoI Government of India GoM Government of Maharashtra GPRS General Packet Radio Service GSM Global System for Mobile Communications HD Hydrodynamic HIS Hydrological Information system HP-II Hydrology Project Phase II IBRD International Bank for Reconstruction and Development IMD Indian Meteorological Department KRBA River Basin Agency MERI Maharashtra Engineering Research Institute MKRBA Maharashtra Basin Agency MODIS Moderate Resolution Imaging Spectro-radiometer MoWR Ministry of Water Resources NIH National Institute of Hydrology, Roorkee NCMRWF National Centre for Medium Range Weather Forecasting NRSA National Remote Sensing Organisation NWP Numerical Weather Prediction QA Quality Assurance QAP Quality Assurance Plan QC Quality Control QPF Quantitative Precipitation Forecast RMC Regional Meteorological Centre (of IMD) RMSE Root Mean Square Error ROS Reservoir Operation System

Interim Report i Krishna & Bhima River Basins RTSF & ROS

RR Rainfall-Runoff RS Remote Sensing RTDAS Real Time Data Acquisition System RTDSS Real Time Decision Support System RTSF Real Time Streamflow Forecasting SAR Synthetic Aperture Radar SO Structure Operation SRTM Shuttle Radar Topography Mission TKRBA Tapi Khandesh River Basin Agency VRBA Vidarbha River Basin Agency WALMI Water and Land Management Institute WB World Bank WRD Water Resources Department

ii Interim Report Krishna & Bhima River Basins RTSF & ROS

Table of Contents

List of Acronyms and Abbreviations ...... i

1 EXECUTIVE SUMMARY ...... V

2 INTRODUCTION ...... 1 2.1 Features of RTSF & ROS ...... 1 2.2 Database-Models Integration ...... 1 2.3 Interim Report ...... 2

3 PROGRESS REPORT ...... 3

4 DATABASE DEVELOPMENT ...... 7 4.1 Data Types...... 7 4.2 Historical Time Series Data...... 7 4.3 Irrigation data ...... 8 4.4 Real Time Data ...... 9 4.5 Static and Annual Data ...... 10 4.6 Database Software ...... 11 4.7 Database Design ...... 13 4.8 Communications ...... 19 4.9 Database Management ...... 19

5 MATHEMATICS MODELS ...... 21 5.1 Introduction ...... 21 5.2 Modelling Software ...... 21 5.3 Hydrology ...... 21 5.4 River Basin...... 23 5.5 River and Flood Plain ...... 24 5.6 Model Integration ...... 25

6 RAINFALL-RUNOFF MODEL ...... 27 6.1 Model Setup ...... 27 6.2 Model Outputs ...... 39 6.3 Calibration ...... 39 6.4 Further Work ...... 48

7 RIVER BASIN SIMULATION MODEL ...... 49 7.1 Model Setup ...... 49

Interim Report iii Krishna & Bhima River Basins RTSF & ROS

7.2 Model Outputs ...... 52 7.3 Further Work ...... 53

8 HYDRODYNAMIC MODEL ...... 55 8.1 Introduction ...... 55 8.2 Model Setup ...... 55 8.3 Calibration ...... 74 8.4 Reservoir Operation ...... 76 8.5 Further Work ...... 79

9 FORECASTING MODEL...... 80 9.1 Introduction ...... 80 9.2 Overview of the Forecasting and Operation System ...... 82 9.3 System Configuration ...... 84 9.4 On-line Operation ...... 85 9.5 Reservoir Operation ...... 88 9.6 Scenario Management ...... 89 9.7 Communication WEB Portal ...... 91

10 CAPACITY BUILDING ...... 95

11 FURTHER WORK ...... 96 11.1 Data Related work ...... 97 11.2 Updating & further development of the models ...... 97 11.3 Implementation of the RTSF & ROS ...... 97 11.4 Support to Capacity Development ...... 97

12 REFERENCES ...... 99

APPENDIX ...... 100 A.1 Sample Satellite Image ...... 101 A.2 Status of Reservoir Data ...... 101 A.3 Range analysis of Data (at Proposed RT stations) ...... 103 A.4 Updated Meta Data (in Separate Volume) ...... 109

iv Interim Report Krishna & Bhima River Basins RTSF & ROS

1 EXECUTIVE SUMMARY

The Project “Consultancy services for the implementation of streamflow forecasting and reservoir operations for Krishna and Bhima River Basins in Maharashtra” commenced with the opening of the project office in Pune on the auspicious day of Ganesh Chaturthi on 17th August 2011. DHI (India) Water and Environment are the Consultants assigned by the Water Resources Department of Government of Maharasthra, India. The assignment is scheduled to be completed in 18 months with an extended technical support period of two years. This Interim Report, which is part of the consultancy project deliverables, presents the progress made during the first eight months of the project. The Report contains a draft knowledge base management system, development of flow and flood forecasting models at pilot levels, the overall reservoir operational guidance system at conceptual stage and a prototype of an internet based communication system. As stipulated in the Terms of Reference (TOR), the interim deliverable are initial demos/models of the knowledge base, real time streamflow/flood forecasting system, reservoir operational guidance systems, and communication and information systems. The initial demos/models were presented in the Interim Workshop held on March 27, 2012 at the YASHADA Centre, Pune. The presentations and demos included:  Draft Knowledge Base: Containing samples of available data relating to GIS data, topographic data, satellite imageries showing administrative/land use/land cover/cropped and irrigated areas, soils, climate, historical hydro-meteorological data, water levels and flow, water resources including reservoirs, facilities for the generation of daily crop water requirements. The knowledge base also has the capability of analysing historical hydro-climatic time series data.  Rainfall-Runoff Model (NAM): Rainfall-Runoff Model for the entire Krishna and Bhima River Basins in Maharashtra calibrated with historical data. The rainfall- runoff model provides inputs in the form of catchment outflows to both the river basin simulation model and the hydrodynamic model.  River Basin Simulation Model (MIKE BASIN): River Basin Simulation Model for the entire Krishna and Bhima River Basins in Maharashtra.  Hydrodynamic Model (MIKE 11): A flow routing model with fully dynamic simulation of river and reservoir system in the entire Krishna and Bhima River basins in Maharashtra. However, this model is at a pilot level due to availability of only a limited number of river cross sections.  Forecasting Model: MIKE 11 Hydrodynamic model for real time forecasting of stream flow with data assimilation for real time updating required for forecasting (operated in a hindcast mode).  Reservoir Operation System: based on the hydrodynamic model, a reservoir operation system integrating all the interconnected reservoirs.  Communication & Information System: A prototype Web Portal to provide access to rainfall, river and reservoir status for the Krishna and Bhima River Basins and forecasts (in a hindcast mode) and reservoir operational scenarios.

Therefore, the TOR requirements for the interim progress have been fulfilled. The simulation models will be further developed with additional data required for improved calibration. The flood forecasting model will only be completed after the river cross section

Interim Report v Krishna & Bhima River Basins RTSF & ROS data are available from the planned field surveys. The forecasting and operational systems will be tested with real time data expected to be available from June 2012 from the on-going RTDAS Project. The capacity building activities, especially training of WRD/BSD officers in modelling, will now be able to use the developed models and they will be involved in the further development, testing and operation of the forecasting system. The database and models and the forecasting system together with computer hardware will be transferred to the Operational Control Room in Sinchan Bhavan, Pune as soon as the physical infrastructure is ready.

vi Interim Report Krishna & Bhima River Basins RTSF & ROS

2 INTRODUCTION 2.1 Features of RTSF & ROS The specific objective of the Project is to develop a Real Time Streamflow Forecasting and Reservoir Operation System (RTSF & ROS) for the Krishna and Bhima River Basins in Maharashtra. The System will integrate the real time Data Acquisition System (RTDAS) with data from external sources, meteorological forecasts, flow forecast modelling, analysis and decision support tools in an IT system designed for ease of use by operators. The main features of the RTSF & ROS are:  Comprehensive database  Comprehensive facilities for integrated presentation of the dynamics of the hydrology and water resources of the basin  A range of hydrological, river basin water resources and hydrodynamic river models  Predictions of the future hydrologic state of the catchment and river system  Reservoir Operation Guidance system The domain of the models of the RTSF & ROS is the area of the Krishna and Bhima River Basins in Maharashtra. 2.2 Database-Models Integration At the core of the RTDSS are mathematical models which describe the state of the catchment and main rivers, and predict future states for a range of scenarios relating to natural events and human intervention. The models require data:  describing the physical features of the catchments, rivers, reservoirs and other hydraulic structures  hydrologic data describing the state of the catchment and rivers – historical data for model calibration  real time and forecast data for making forecasts of future catchment states  water demand data for optimizing the operation of reservoirs All data used for modelling purposes and output from model simulations are stored and maintained in the database. The RTSF & ROS provides a large number of functionalities for working with data, comprising database input and output tools, data visualisation and data processing (filtering, gap filling, etc). Modelling-wise the RTSF & ROS includes functionality for automatically extracting and arranging the necessary data for the model simulations and subsequently for importing the generated data to the database. This ensures that data (covering both observations and model output) are readily available in the RTSF &ROS user interface and that system management becomes easier compared to having data stored in file system folders.

Interim Report 1 Krishna & Bhima River Basins RTSF & ROS 2.3 Interim Report This Interim Report which is part of the consultancy project deliverables, contains a draft knowledge base management system, development of flow and flood forecasting models at pilot levels, a reservoir operational system also at a pilot level and a prototype of an internet based communication system. The report presents the design of the database and management system as well as presents the data available to the project so far. Details of the database are presented in the Appendix. The database will be updated as all data becomes available including the real time data from the RTDAS. Similarly, the models will be finalised after all the required data are available and an operational system will be developed and implemented after regular real time data start flowing from the RTDAS. The details of the updated database will be presented in the next interim report. Detailed description of model updates and the RTSF & ROS, including user guides and other documentation will be presented in the Draft Final Report.

2 Interim Report Krishna & Bhima River Basins RTSF & ROS

3 PROGRESS REPORT

Table 3.1 presents a summary of tasks and sub-tasks completed. The tasks refer to those stipulated in the Contract, which were expected to be completed and the outputs included in the Interim report. The Contract also stipulates that the Interim Report activities should include initial demos/models of the knowledge base, real time streamflow/flood forecasting system, reservoir operational guidance systems, and communication and information system. The initial demos/models were demonstrated at the Interim Workshop held on 27 March 2012. It was seen that the stakeholders appreciated the models and demonstrations. However, it was recognised that adequate data required for completion of the models with reasonable accuracy is needed to be acquired. Also, in order to implement the forecasting and reservoir operation system in real time, the RTDAS project should be completed well in time. Figure 3-1. presents a glimpse of the progress in Microsoft Project Manager.

Interim Report 3 Krishna & Bhima River Basins RTSF & ROS

Figure 3-1 Progress Tracking in MS Project Manager

4 Interim Report Krishna & Bhima River Basins RTSF & ROS

Table 3.1 Summary of Progress of Stipulated Tasks Task Sub-tasks/activities Progress Outputs Task 1 Six sub-tasks related to review and Completed in the inception Inception Report planning of activities, development of phase (17 Aug – 16 Nov, Review Current modelling framework and training plan 2011) Forecasting and Operational Capabilities (Inception Phase) Task 2 (2.1) Functional specifications for the A knowledge base system designed Knowledge base containing all WRD Krishna-Bhima knowledge base and installed: all historical hydro- available time series data, GIS Knowledge Base meteorological data, river flows and data, available satellite Development (2.2) Design and develop database levels, irrigation data, available images, available data of management system satellite images and other GIS data reservoirs. (2.3) Develop knowledge base collected and populated in the database. Data from most of the reservoirs collected and included in the database. Collection of remaining data on-going. For the real time data, links to be established with RTDAS, IMD and others. Task 3 (3.1)Based on the modelling framework Progress on development of pilot A Rainfall-runoff model (NAM) set out in Task 1, the modelling system models as anticipated considering for the Krishna and Bhima Real-Time Streamflow / will be established and calibrated against the data availability. The rainfall- River Basins developed with Flood Forecasting Model historical and current data runoff model is fairly well calibrated 122 catchments against historical data. The A River Basin simulation calibration of hydraulic model and model (MIKEBASIN) hence the forecasting and reservoir developed for the entire basin operation system is at a pilot level (Krishna-Bhima) awaiting detailed river cross A Hydrodynamic model

Interim Report 5 Krishna & Bhima River Basins RTSF & ROS

Task Sub-tasks/activities Progress Outputs sections. (MIKE11) for the two basins including all rivers and A Web-based real time reservoirs developed at a pilot communication system with a level using available river capability of showing status of the cross section data reservoirs, rivers, rainfall and forecasts in a hindcast mode is A forecasting system prepared at demonstration level. developed and demonstrated hindcasting historical data and events. An example of reservoir operation guidance system developed and demonstrated An internet based communication system (WEB Portal) developed and demonstrated for dissemination of status and forecast of hydrological and hydraulic parameters of the system.

Task 6 Continuous activity including institutional On-going: basic and introductory A high performance database strengthening, training and workshops, training on hydraulics, hydrological server. Capacity Building and study tours, etc. and hydrodynamic modelling carried Training out. Database server procured for the Operational Control Room.

6 Interim Report Krishna & Bhima River Basins RTSF & ROS

4 DATABASE DEVELOPMENT 4.1 Data Types Data types may be classified according to the frequency at which the data change, and also reflect the means of data collection:  Time Series – comprise point based and raster based measurements of meteorological and hydrometric data made at variable intervals.  Annual Data – comprise surveys updated annually or every few years, such as river cross sections and reservoir bathymetry. Water level-discharge rating curves are also included in this category.  Static Data – comprise data which do not normally change over time, such as structure geometry and basin topography. Rating curves for gates and power stations are included in this category. The availability of these data is discussed in the following sections. 4.2 Historical Time Series Data The historical time series data include rainfall, evaporation, full climate, discharge, water level, reservoir inflow outflow, irrigation demand and supply and other water supply data. Table 4.1 shows a summary of historical time series data collected and included in the database. Details of the data (Meta Data) are given in Appendix A-4.

Interim Report 7 Krishna & Bhima River Basins RTSF & ROS

Table 4.1 Summary of historical time series data Data type No. of Frequency Years of data stations

Rainfall 185 Daily (185) Varying from station to station (mostly 1971- Hourly (76) 2009)

Evaporation 44 Twice Daily Varying from station to station (mostly 1999- 2009)

Full Climatic 42 Hourly (date Varying from station to (temperature, from 17 station (mostly 2002- humidity, wind, stations 2009) sunshine, received) rainfall)

Discharge 25 Daily (25) Varying from station to station (mostly 1991- Hourly (21) 2009)

Water level 25 Daily (25) Varying from station to (river) station (mostly 1971- Hourly (21) 2009)

Reservoir 40 daily Varying from station to inflow-out flow station (mostly 2002- 2009)

4.2.1 Satellite Data A variety of satellite data are to be stored in the database. Sample satellite images are shown in Appendix A. 4.3 Irrigation data A tool has been developed to compute daily irrigation requirement for different crops and for different areas using the data from Full Climate Stations (FCS). The FAO56 method of calculation of crop water requirement has been adopted as a standard tool in the MIKEBASIN Software. This will enable computation of daily irrigation requirement in real time whenever the on-going RTDAS project will provide real time data from the FCS. Figure 4-1 shows an example of daily irrigation data calculated using the FAO CROP WAT tool in MIKEBASIN.

8 Interim Report Krishna & Bhima River Basins RTSF & ROS

Figure 4-1 Computation of Daily Irrigation Requirement 4.4 Real Time Data

4.4.1 RTDAS A comprehensive range of high quality ground based point measured data describing the state of the basin and the rivers and reservoirs will become available in real time with implementation of the Real Time Data Acquisition System (RTDAS) from June 2012. The real time network includes rainfall, climatic, reservoir water level and discharge data. WRD has entered into a contract with Mechatronics for the installation of a network of ground based measuring stations recording a variety of parameters, transmitting the data in real time via INSAT and GPRS to a receiving station at the RTDAS data Centre located in Sinchan Bhaban, Pune. A summary of the stations and data types is presented in Table 4.2.

Table 4.2 List of Real Time Data from the RTDAS (on-going) Data type Number of stations Remarks Rainfall 149 Full Climate Station 42 Water Level (river) 31 Discharge (river) 31 Water Level (reservoir) 46 59 gate openings (in selected dams)

4.4.2 Meteorological Forecasts It is proposed that WRD sign a Memorandum of Understanding with the Indian Meteorological Department (IMD) and the National Centre for Medium Range Weather Forecasting (NCMRWF), for the provision of rainfall forecasts. Also forecasts of rainfall and other weather parameters will be downloaded on a daily basis from the IMD’s website (www.imd.gov.in). The IMD website also provides a link to satellite data. A request is also being made to the Regional Integrated Multi-Hazard Early Warning System (RIMES) www.rimews.org, an inter- governmental organisation based in Bangkok, Thailand to provide high resolution quantitative precipitation forecasts.

Interim Report 9 Krishna & Bhima River Basins RTSF & ROS 4.5 Static and Annual Data

4.5.1 Rating Curves from Gauging Stations Rating curves from discharge gauging stations are prepared by WRD, generally at the end of each monsoon season. The most recent rating curves are available for all discharge sites. Historical rating curves are also available. The rating curves are used to compute the discharges from the water level measurements. 4.5.2 River Cross Sections River cross sections from different sources are available (presented in Chapter 8) and a large number of river cross sections are planned to be surveyed by WRD. 4.5.3 Reservoir Bathymetry Area-elevation and volume-elevation curves are available for 31 reservoirs (out of 46) in the two basins. These have been computed from the topographic survey maps at the time of reservoir construction. Appendix A presents the status of reservoir data.

4.5.4 Digital Elevation Model A digital elevation model of the entire basin has been obtained from the NASA SRTM, at resolutions of 30m and 90m. The DEM, together with the catchments and rivers derived from the DEM, is illustrated in Figure 4-2. Each grid square represents the average elevation within the square, with an accuracy in the order of ±20m.

Figure 4-2 Digital Elevation Model (SRTM)

10 Interim Report Krishna & Bhima River Basins RTSF&ROS

4.6 Database Software

4.6.1 Introduction The database will collect and store a wide range of historical and real time hydrology related data on the basins and from external sources, and provide access to RTSF & ROS operators, and on-line and remote web users. In addition to providing the input data for the mathematical models, the database will also store the results from the models. Proprietary software is required for data storage and data display. The database will be used to store historical hydrologic data on the basin and data collected through the RTDAS, definitions of the various scenarios that WRD will utilise for short and long term planning, and input that can be used to operate the dams and other controls. Data with a spatial context will be accessed in a GIS environment. DHI and Mechatronics will agree on the exchange protocol and format of data to be delivered by the RTDAS to the RTSF & ROS. RTDAS will carry out quality control and present the data to RTSF & ROS. The RTSF & ROS system will at predefined frequencies import newly received and checked data to the Database. The requirements for the database can be grouped by:  Enterprise abilities  Manageability  Functionality  Operating system Although there exists a large number of both commercial and open source based database systems, only a limited number of these are applicable to a complex system such as the Krishna-Bhima RTSF & ROS. The basic requirements are that the database is based on both a mature and a widely supported technology. After reviewing many candidate database systems for evaluation with respect to the requirements, the system used in the DSS (Planning) Project and the BBMB RTDSS project has been adopted. The proposed system has a long development history, proven high availability capabilities, a broad user community, and support GIS data. The PostgreSQL together with PostGIS has been adopted as the database solution for the RTSF & ROS. 4.6.2 Software

Database The database component is a relational database management system (RDBMS) storing data in the form of related tables. Relational databases require few assumptions about how data is related or how it will be extracted from the database. As a result, the same database can be viewed in many different ways. The RDBMS is prepared for handling all types of DSS data: GIS (spatial) data, time series data and scenario/model data. The Database components used in the system comprise:  PostgreSQL – a standard well-proven Enterprise-level RDBMS

Interim Report 11 Krishna & Bhima River Basins RTSF & ROS

 PostGIS – an extension to PostgreSQL that makes it possible to maintain and process GIS data

PostgreSQL PostgreSQL is an object-relational database management system. It is released under a Berkeley Software Distribution (BSD) style license and is thus free and open source software. As with many other open source programs, PostgreSQL is not controlled by any single company, but has a global community of developers and companies to develop it. The development of PostgreSQL dates back to the early 1980s. The main features of PostgreSQL comprise:  Stored procedures can be written in high-level languages like Python, C++ and Java  Indexes – based both on column values and expressions. Partial indexes are also supported  Triggers can also be coded in high-level languages  Multi-version concurrency control which provides individual user snapshots of the database  Updatable views  A wide variety of data types  User defined objects  Inheritance – tables can inherit characteristics from a parent table. This can be used to implement table partitioning The PostgreSQL database is described in more detail on the PostgreSQL home page (http://www.postgresql.org). The DSS uses PostgreSQL to store all data as shown in Figure 4-3. The input to PostgreSQL is SQL statements and the output is result sets from the executed SQL statements.

PostgreSQL

Get Get User DSS Front-end DSS Database Put Put

Insert/ Update/ Load Delete

Database tables

Figure 4-3 Database Dataflow

12 Interim Report Krishna & Bhima River Basins RTSF&ROS

PostGIS PostGIS is an open source software program that adds support for geographic objects to the PostgreSQL object-relational database. PostGIS follows the Simple Features for SQL specification from the Open Geospatial Consortium. The first version of PostGIS was released in 2001. The main features of PostGIS comprise:  Geometry types for points, linestrings, polygons, multipoints, multilinestrings, multipolygons and geometrycollections  Spatial predicates for determining the interactions of geometries using the 3x3 Egenhofer matrix (provided by the GEOS software library)  Spatial operators for determining geospatial measurements like area, distance, length and perimeter  Spatial operators for determining geospatial set operations, like union, difference, symmetric difference and buffers (provided by GEOS)  R-tree-over-GiST (Generalised Search Tree) spatial indexes for high speed spatial querying  Index selectivity support, to provide high performance query plans for mixed spatial/non-spatial queries  Raster data in the form of ASCII grids and GeoTiffs as gridded rasters and geo-referenced images (BMP, JPG, GIF, PNG). 4.7 Database Design

4.7.1 Introduction The Krishna-Bhima Database is a tailored database system developed using the above described software. Figure 4-4 shows the overall contents of the Krishna- Bhima database system.

Figure 4-4 Overall Content of the RTSF & ROS Database Management System

Interim Report 13 Krishna & Bhima River Basins RTSF & ROS

4.7.2 Data categories The database stores a wide range of data. The data are categorised according to the format in which they are stored. The link between the data types, which essentially describes how the data are collected, and the data categories is set out in Table 4.4. Table 4.4: Data Categories

CATEGORY FORMAT TYPES

Spatial Data Shapefile DEM Image Remote Sensing Grid Meteorological Forecasts Temporal Time Series Ground Based Point Data Data (historical) Remote Sensing RTDAS Meteorological Forecasts Numerical Rainfall-Runoff Model Parameter Models (NAM) Files Water Resources Reservoir Scenario (MIKE Basin) Bathymetry Definitions Hydrodynamics River Cross Sections (MIKE 11) Structure Geometry

The shapefile format is the most common file format for storing spatial related information. The format is developed by Environmental System Research Institute (ERSI) and is an open and well defined format supported by most providers of spatial information. Some data types appear in both spatial and temporal data categories, ie remote sensing and meteorological forecasts. Figure 4-5 shows the folder structure of GIS data.

14 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 4-5 Folder Structure of GIS Data

4.7.3 Historical and Real Time Series Historical time series data have been imported to the database and organized in the folder structure shown in Figure 4-6. Real time data from the RTDAS will be stored in the database in a similar folder structure. The exact structure will be defined once the RTDAS is defined and installed. All data quality check will take place in the RTDAS. When data becomes available in RTDAS it will be automatically transferred and stored in the RTDSS database. A status message will be parsed for each station every time a new set of measurements is received from the RTDAS.

Interim Report 15 Krishna & Bhima River Basins RTSF & ROS

Figure 4-6: Historical Time Series

4.7.4 Analysis of Time Series Data The Time series manager of the RTDSS Database provides management and analysis functionalities for storing, querying, importing, exporting and quality checking. In addition, the manager offers a suite of tools for visualising, statistical analysis and processing one or more time series. It may also be used to view data prior to model simulations and to analyse and visualise model simulation outputs. Typical uses of the time series manager are listed in Table 4.5. Table 4.5 Examples of the time series manager Task Principal activities Create Time Series A new time series can be created in the database by: directly creating a series using the tool, importing a time series from a variety of sources including RTDAS. Edit Time Series An existing time series can be edited for all its components (time, value) Export Time Series A time series may be exported to Excel, a modelling system

16 Interim Report Krishna & Bhima River Basins RTSF&ROS

or an external file to a prescribed format. Filter Time Series This function is very useful for large databases. Defining a search criteria for looking-up time series in the database, such as name, type data, scenarios etc. Import Time Series Importing Time series data from Excel, SWDES etc. Inspect Time Series Looking at the time series attributes and meta data Visualize Time Activities include display time series data in tabular or Series graphical forms, adding time series to an existing series or chart, customizing the chart or table Using/Processing This functionality is the most useful in data processing, Time Series quality checking, gap analysis, resampling, statistical analysis, etc.

In order to illustrate the time series analysis capabilities of the database, rainfall data form and discharge data from Karad G-D station were selected. Figures 4-7 to 4-9 Show the performed analyses. Other detailed analyses including identification of average, minimum and maximum values are presented in Appendix A.

Figure 4-7 The Daily Rainfall at Mahabaleshwar (top graph) has been re-sampled into monthly (middle graph) and yearly (bottom graph).

Interim Report 17 Krishna & Bhima River Basins RTSF & ROS

Figure 4-8 Discharge data at Karad G-D Station (top graph) has been used for duration analysis (middle graph) and statistical analysis (seasonal annual maximum)

Figure 4-9 Cumulative probability distribution function (CDF) of discharge at Karad G-D Station

18 Interim Report Krishna & Bhima River Basins RTSF&ROS 4.8 Communications

4.8.1 External Data Three types of external data source will be interfaced to the Krishna-Bhima RTDSS database:  Data from the Real Time Data Acquisition System (RTDAS)  Data derived from Satellite images  Weather data including forecasts e.g. from IMD, RIMES etc.

Data from Satellite Images Satellite images are to be stored in a folder structure outside the DSS database in Environmental System Research Institute (ESRI) .asc GRID format. Data derived from the Satellite images will be stored in the DSS database as Time series data.

4.8.2 Dissemination It is proposed to install the database system at six offices: BSD, Operational Control Room (Sinchan Bhavan Pune), Office of the Chief Engineer (Hydrology and Planning) at Nashik, and offices of three Chief Engineers at the Sinchan Bhavan. Password protection (as shown in Figure 4-9) may be implemented as per the direction of the Chief Engineer (Planning & Hydrology). In addition, relevant WRD Offices will have access to all relevant data from the RTDSS Centre database through a Web based application.

Figure 4-9 Password Protection of the RTDSS database

4.9 Database Management

4.9.1 Installation The following software packages must be installed before the installation of the Krishna-Bhima RTDSS can proceed:

Interim Report 19 Krishna & Bhima River Basins RTSF & ROS

(1) The Krishna-Bhima RTDSS is a software package which builds upon Microsoft .Net Framework 3.5 SP1 and thus depends on it being installed for the correct functionality. (2) The RTDSS uses PostgreSQL 9 as the database server and PostGIS 2.0 as the GIS extension to PostgreSQL. (3) The DHI modelling tools MIKE Basin and MIKE 11 need to be installed to run simulations. The current version of these modelling tools is 2012. MIKE Basin is an extension to ArcMap and thus requires ArcMap to be installed. Required version is ArcMap 10. The modelling tools are only required for model calculations. Other use of the system does not require the modelling tools to be present. The main copy of the RTDSS Database will be installed in the main server at the Operational Control Room. 4.9.2 Staff Training System Administration corresponds to the skills associated with administering the RTDSS system. An IT personnel shall be required to work as (a) Computer Systems Manager, (b) Network Manager, (c) Data Manager, and (d) Website administrator. The IT personnel has to receive specialised training on the above areas. A hands- on and on-the-job training will be provided by the Consultants. It is also expected that the RTDAS contractor will provide adequate training to WRD staff during the installation and maintenance periods.

4.9.3 Operation and Maintenance Most database administration tasks can be performed from within the RTDSS, eg user management, data import and data export. There are a few situations where a systems administrator must interact directly with the database, from outside the RTSF & ROS System. Examples comprise backup and restore. It is essential that the database server is secured by a well functioning backup solution. The database should have an incremental backup daily and a full backup weekly. Integration of the backup tools with the backup system should be handled by the Client’s System Administration organisation.

20 Interim Report Krishna & Bhima River Basins RTSF&ROS

5 MATHEMATICS MODELS 5.1 Introduction The core function of the mathematical models in the RTSF & ROS is the processing of the real time and historical data from the basin into quantitative short and longer forecasts of the inflows to the reservoirs, and flood forecasts together with the simulation of the possible consequences of various options for operating the reservoirs. The model simulations will run extracting real time data from the database and producing forecasts following fixed time schedules. More advanced analyses, typically with uncertain input, will be carried out as what-if scenarios, revealing possible consequences of the assumed inputs of meteorology and operation scenarios. The domain of the models of the RTSF & ROS is the area of the Krishna and Bhima River Basins in Maharashtra. The models will supplement the existing decision making in WRD, which to a large extent is based on decades of experience operating the dams, and on historical data and derived statistics. This Section describes the models developed so far in the project. The models are based on historical data available so far and on available river cross sections and reservoir information. The Final Model Development Report will be prepared after all data becomes available including the real time data from the RTDAS. The models presented may be considered prototypes. While this report presents the establishment and functions of the basic models, a final report will be prepared with complete documentation after implementing the real time forecasting models within the RTSF & ROS. Final establishment of the models and the RTSF & ROS will only be done through the testing and operation procedures in conjunction with the RTDAS, expected to commence in June 2012. During that phase, calibration of the model and forecasting parameters will be refined with comprehensive high quality data from the RTDAS. Reservoir operational procedures including forecasting, scenarios and optimisation will be defined and fully implemented. Flood forecasting and floodplain mapping will be carried out after incorporating the cross section data to be obtained from the proposed field surveys. 5.2 Modelling Software The mathematical models describe the hydrologic state of the catchment and main rivers, and predict future states for a range of scenarios relating to natural events and human intervention. Compatible, well tested proprietary software products with open scripting interfaces to access input and output, and corresponding hardware, are required to run the model simulations in the shortest possible time, while providing the required level of detail and accuracy. The MIKE software system, developed by DHI Water Environment Health, based primarily on the need for advanced data assimilation for optimal flood forecasting, options for reservoir operation, has been adopted for this project. This package fulfils the entire features and functionality required for the Krishna-Bhima RTSF & ROS. 5.3 Hydrology Hydrological modelling is applied to simulate the rainfall-runoff process in order to predict the runoff from gauged and un-gauged catchments in the basin.

Interim Report 21 Krishna & Bhima River Basins RTSF & ROS

The NAM model is a hydrological model of the continuous moisture accounting type allowing for both long term water balance and runoff estimates, and for short term inflow forecasting. In essence the NAM model simulates the runoff from the various sub-catchments in the system on the basis of an input of historical and predicted precipitation and potential evapotranspiration (Figure 5-1). The NAM model is coupled with the other MIKE models forming part of the integrated modelling system built into the RTDSS.

Figure 5-1: Structure of NAM Model For more specific information of the NAM modelling system please refer to the supplier’s extensive Reference and User Guides. Below is the list of NAM parameters with their functions. Umax Max. water content in surface storage (mm) Lmax Max. water content in root zone (mm) CQOF Overland flow runoff coefficient CKIF Time constant for routing interflow (hour) CK1,2 Time constant for routing overland flow (hour) TOF Rootzone threshold value for overland flow TIF Rootzone threshold value for interflow TG Rootzone threshold values for ground water recharge (m) CKBF Time constant for routing baseflow (hour) Carea Ratio of Ground water area to catchment area Sy Specific yield of ground water reservoir GWLBF0 Threshold Ground water depth for baseflow GWLBF1 Capillary flux (depth per unit flux) Cqlow Lower baseflow, recharge to lower reservoir Cklow Time constant for routing lower baseflow (hour)

22 Interim Report Krishna & Bhima River Basins RTSF&ROS

5.4 River Basin MIKE BASIN is a simulation model for water allocation representing the hydrology of the basin in space and time. Technically, it is a network model in which the rivers and their main tributaries are represented by a network of branches and nodes (Figure 5-2). The branches represent individual stream sections while the nodes represent confluences, bifurcations, locations where certain water activities may occur, or important locations where model results are required. MIKE BASIN operates on the basis of a digitised river network generated directly from the digital elevation model in ArcGIS. All information regarding the configuration of the river branch network, location of water users, channels for intakes and outlets to and from water users, reservoirs are also defined by on- screen editing in ArcGIS.

Figure 5-2: Simplified Schematisation of MIKE BASIN Model MIKE BASIN focuses on the allocation of water for various purposes and includes an efficient module for operation of reservoirs and for hydropower simulation. It uses a conceptual rather than a hydrodynamic flow description of the river channels making it particularly well suited for long term forecasting and long term simulation for optimisation of reservoir operations where calculation speed plays an important role. The model has an efficient module for reservoir operations including both normal rule curve reservoirs and allocation pool reservoirs, where the various users operate their allocated share of the reservoir volume independently. Minimum releases for downstream river reaches, flood control, restrictions or extra hydropower releases in cases of water shortage or abundance are easily handled. Basic input to the model consists of time series data of various types. Essentially only time series of catchment runoff is required to have a model setup that runs. In

Interim Report 23 Krishna & Bhima River Basins RTSF & ROS

the RTSF & ROS system these runoff series are simulated discharge series from the hydrological model. Additional input files define reservoir characteristics and operation rules for each reservoir, meteorological time series and hydropower characteristics and flow demands. MIKE BASIN has the ability to allow flexible time stepping from seconds to months and years enabling the user to run the model at the level of detail and with the processes required. It is thus suitable for use in real time and near real time. In simulation mode the output from the model is the flows in the various channels, water levels and volumes in reservoirs, spills and canal releases for irrigation or other water uses, as well as turbine flows and hydropower production, as generated by a certain inflow or release scenario. In optimisation mode the model may be used to generate a certain release series fulfilling certain criteria such as a certain probability of maintaining a predefined cushion at the end of the depletion season to safeguard against a possible delayed or weak monsoon. 5.5 River and Flood Plain The software selected for the river, reservoir and flood plain modelling, MIKE 11, provides an array of computational methods for steady and unsteady flow in branched and looped channel networks, and flood plains. MIKE 11 can be applied to flow conditions ranging from steep river flows to tidally influenced narrow estuaries, and describes local subcritical and supercritical. MIKE 11 includes advanced formulations for simulating flow through a variety of standard and complex structures, including operational structures and dam break The modules applied to the Krishna-Bhima RTSF & ROS are:  Full hydrodynamics  Structure operations  Inflow and flood forecasting  Autocalibration and optimisation MIKE 11 uses the high order fully hydrodynamic model providing improved numerical stability. MIKE 11 incorporates a wide range of hydraulic control structures, which can be operated on the basis of hydraulic conditions and structure operations upstream and downstream in the basin. MIKE 11 has advanced data assimilation making the maximum use of the real time information from the basin to predict the future state in terms of water levels and discharges. In order to optimise hydraulic control operations, an optimisation routine is required, based on multiple user defined objectives. MIKE 11 has facilities for automatic calibration applied in optimisation. Figure 5-3 shows the modular structure of MIKE 11.

24 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 5-3 Modular Structure of MIKE 11 5.6 Model Integration To allow the users of the RTSF & ROS also called a Real Time Decision Support System (RTDSS) to focus on key decision making, and minimise the computer operation tasks, manual input from the user will be kept to a minimum. The model results are presented in specific plots and tables designed for easy understanding and use of decision makers and other stakeholders. For normal operation, all the models will be integrated within the RTSF & ROS. The complex user interfaces to the individual models giving access to all parameters and the full output of each model is intended to be used only during establishment and recalibration of the models. For normal operations such as pre- scheduled forecasts or established scenarios, the models will be activated either automatically (pre-scheduled execution) or through the common user interface. The interface will be tailor made for the RTSF & ROS, and will include only the input requirement necessary for these tasks. The observed historical data from the basins as well as the currently collected real time data is stored in the Database. The Database also holds one or more copies of the each of the three types of models and information on which data each model needs to access when used in its various applications. Typically, a model task will start with the appropriate model being extracted from the database with all necessary input updated with the latest information from the RTDAS. If input data need to be changed this will be done in the extracted copy of the data. An example could be to analyse the reservoir inflow in the case that the rainfall in the forecasting period would be 20% higher than the prediction already in the database. After simulation, essential model results for further analyses or as input to subsequent model applications is stored in the database from where they are extracted by the subsequent application. If the models are used in optimisation mode, eg to seek optimal releases or spills, the MIKE 11 and MIKE Basin model runs through an optimiser that executes the model a large number of times to find an optimal solution.

Interim Report 25 Krishna & Bhima River Basins RTSF & ROS

A typical application involves:  Input of the necessary scenario parameters. In case of a pre scheduled application this will not be necessary  Run the hydrological model (in case of a long-term forecast this will take place as number of individual one-year simulations for which the results will subsequently be merged.  Run the simulation model (MIKE11 or MIKE Basin)  Optimise the river model (optional)  Present the results. The typical flow of information for a short term forecast is shown in Figure 5-4.

• Update RTDSS Database with weather forecasts, satellite information and RTDAS Update Database • Input reservoir releases and power generation • specify additional scenario input, eg +20% Specify forecast rainfall (optional) • Simulation launch Input • Extract Hydrology Model and Input Data from Database Hydrology • Run Hydrology Model (hotstart) Model • Store results of Hydrology Model • Extract Hydrodynamic Model and Input Data including Hydrology Model results from Database • Perform Data Assimilation on real time discharges Hydrodynamic and run Hydrodynamic Model (hotstart) Model • Store results of Hydrodynamic Model

• Extract and present selected results and information in standard RTDSS web pages Present Results

Figure 5-4: Flow of Information for short term forecast

26 Interim Report

Krishna & Bhima River Basins RTSF&ROS RTSF&ROS 6 RAINFALL-RUNOFF MODEL

6.1 Model Setup

6.1.1 Catchment Delineation To simulate the spatial variation in the lateral inflow to the river system, the two basins have been subdivided into 122 sub-catchments as shown in Figure 6-1. The sub-catchment delineation is to a large degree been based on gauging station location to make it possible to calibrate the model at as many locations as possible. Further sub-catchments have been defined at locations where important tributaries join the main rivers and where spatial variation in precipitation or terrain indicate the need for a subdivision. Table 6.1 shows the area of each sub- catchment. In delineating the 122 sub-catchments the following factors have been considered: topography, rainfall variation, sub-basin outlets, watershed atlas produced by Soil & Land Use Survey of India (www.cgwb.gov.in) and the Maharashtra Water & Irrigation Commission Report (1999). The present delineation of catchments is in agreement with the sub-basin map available in the above report. The All India Soil and Land Use Survey (AISLUS) Organization (now known as Soil and Land Use Survey of India) of the Department of Agriculture and Cooperatives has published a national level watershed atlas on 1: 1 million scale using the base map from irrigation atlas of India in the year 1990. In this atlas, the entire river systems of the country have been divided into 6 Water Resources Regions, which have been further divided into 35 basins and 112 catchments. These catchments have been further divided into 500 sub-catchments and 3237 watersheds. The atlas consists of 17 sheets on 1:1 million scales along with a compendium of watersheds giving details of other related information such as area within the basin, sharing states and stream names etc. This atlas is being extensively used for various purposes by all the State and Central Government agencies, including WRD and GSDA of Government of Maharashtra.

Interim Report 27 Krishna & Bhima River Basins RTSF & ROS

Figure 6-1: Sub-Catchment Delineation of Krishna and Bhima River Basins Showing rainfall stations, evaopration stations, reervoirs and G-D stations

28 Interim Report

Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Table 6.1 NAM Catchment Area

No Catchment Name Area (km2 ) No Catchment Name Area (km2 ) No Catchment Name Area (km2 ) 1 PUSHPAVATI_R1 104.0 41 BHOGVATI_R1 132.20 82 KOYNA_R2 261.20 2 PUSHPAVATI_R2 90.5 42 SINA_R1 1474.69 83 KRISHNA_R4 239.16 3 PUSHPAVATI_R3 106.2 43 SINA_R2 4154.32 84 KRISHNA_R5 31.29 4 GHOD_R1 273.7 44 PAWANA_R2 38.96 85 KRISHNA_R6 710.50 5 BHIMA_R1 45.0 45 INDRAYANI_R3 30.58 86 KRISHNA_R7 1779.12 6 BHIMA_R2 254.2 46 INDRAYANI_R4 21.02 87 YERALA_R2 1317.44 7 KUKADI_R1 369.4 47 INDRAYANI_R5 440.59 88 KRISHNA_R8 177.11 8 ANDHRA_R1 124.1 48 MUTHA_R2 46.85 89 KRISHNA_R9 440.32 9 INDRAYANI_R1 45.6 49 MULA_R2 85.85 90 KRISHNA_R10 21.78 10 BHAMA_R1 191.9 50 MUTHA_R3 225.12 91 KASARI_R2 552.37 11 ANDHRA_R2 90.7 51 TARALI_R2 55.16 92 KUMBHI_R2 98.71 12 INDRAYANI_R2 14.7 52 MAN_R-K 39.86 93 DHAMNI_R2 110.98 13 PAWANA_R1 116.6 53 WANG_R1 69.90 94 BHOGVATI_R2 640.06 14 MULA_R1 248.9 54 MUTHA_R4 124.39 95 PANCHAGANGA_R1 261.56 15 MUTHA_R1 36.0 55 PAWANA_R3 310.06 96 PANCHAGANGA_R2 427.67 16 MOSE_R1 131.6 56 BHIMA_R3 85.66 97 KRISHNA_R12 170.65 17 AMBI_R1 118.2 57 BHAMA_R2 37.79 98 DUDHGANGA_R2 400.16 18 GHOD_R2 1071.7 58 MULA_R3 598.10 99 AGRANI_R1 480.54 19 KUKADI_R2 923.1 59 MULAMUTHA_R1 638.83 100 AGRANI_R2 549.18 20 GHOD_R3 546.7 60 BHIMA_R6 1005.03 101 MAN_R-B 4614.47 21 GHOD_R4 912.3 61 MULAMUTHA_R2 243.34 102 BHIMA_R11 1259.43 22 KANAND_R1 51.5 62 BHIMA_R7 678.17 103 SINA_R4 2997.37 23 NIRA_R1 111.0 63 BHIMA_R8 3620.71 104 BHIMA_R12 650.49 24 NIRA_R2 273.6 64 GUNJAVANI_R1 41.97 105 MINA_R1 133.69 25 NIRA_R3 131.1 65 KARHA_R1 397.06 106 BHIMA_R4 735.45 26 KRISHNA_R1 39.8 66 KANAND_R2 432.14 107 BHIMA_R5 66.24 27 KRISHNA_R2 155.5 67 NIRA_R4 709.58 108 PANCHAGANGA_R3 113.80 28 MAHU_R1 28.9 68 KARHA_R2 734.76 109 KRISHNA_R11 139.21 29 KOYNA_R1 897.9 69 NIRA_R5 1243.24 110 WARANA_R3 248.86 30 VEENA_R1 212.3 70 BHIMA_R9 1929.64 111 BORI_R1 1628.40 31 URMODI_R1 112.7 71 NIRA_R6 486.33 112 KOYNA_R3 1165.50 32 TARALI_R1 80.5 72 BHIMA_R10 3346.74 113 NIRA_R7 1047.92 33 MORNA_R1 54.3 73 KRISHNA_R3 837.00 114 BORI_R2 539.94 34 WARANA_R1 274.1 74 VASANA_R1 526.95 115 SINA_R3 3636.28 35 KADAVI_R1 24.0 75 URMODI_R2 22.20 116 Chincholi_R1 876.23 36 KASARI_R1 32.7 76 KERA_R1 107.43 117 Agrani_R3 549.50 37 KUMBHI_R1 20.8 77 MORNA_R2 60.56 118 Krishna_R13 972.87 38 DHAMNI_R1 42.6 78 WANG_R2 162.95 119 Vedganga_R1 767.35 39 TULSHI_R1 34.3 79 KADAVI_R2 330.33 120 HiranyKesi_R1 230.30 40 DUDHGANGA_R1 203.5 80 YERALA_R1 766.31 121 Ghatprabha_R1 343.20 81 WARANA_R2 1151.37 122 Tamraprani_R1 30.25 Note: Catchment numbers 116 to 122 are not calibrated yet (ungagged catchments).

Interim Report 29 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

6.1.2 Model Input Data The driving input to the NAM model is time series of precipitation and potential evaporation. The input time series must cover the entire period to be simulated. The model includes facilities for weighted averaging and gap-filling of single station records to produce representative mean areal series for each catchment. In addition to these driving inputs, recorded flows and snow coverage are used for calibration and for real time data assimilation.

Precipitation Precipitation is the input to the hydrological model with the largest impact on the generated flows. Two sources of precipitation data are available for the establishment of the model, namely:  Daily time series from 109 rainfall stations in the basins (2000 – 2009)  Hourly time series from 76 stations (2000-2009) It was decided to use those rainfall stations for the NAM modelling that had at least 9 years long record between 2000 and 2009. In that way 155 stations were found suitable for the calibration of the NAM model. A shape file of these stations including the annual average rainfall amount as an attribute was also prepared (Figure 6-2). This was very useful in the later stage as the annual rainfall map for the basin was not available. In the mountainous area the average annual rainfall can vary by 3 to 4 times in a distance of only 20 kilometres. This was considered while doing the calibration of the catchments if the Thiessen weighting did not work. The Nam model determines the weighted average rainfall over and evaporation from each catchment by the Thiessen Polygon method. The Thiessen polygons were prepared for the set of 155 rainfall stations and weights were derived for each of the catchments. If there were less than three rainfall stations per catchment additional stations were chosen for calculation of the catchment rainfall. The lowest weight was chosen as 0.05 and all the weights have two decimal places. In case of more than three stations per catchment the stations with weight values of less than 5 % were not taken into account. The correction factors due to the unequal rainfall distribution were derived during the process of calibration. In the mountainous areas the Thiessen weighting simply did not work due to the huge spatial variability of the rainfall. On a distance of 10-20 km the average annual rainfall can vary from 2 to 3 times. In the absence of the rainfall map for the considered region the initial weights were corrected by assuming a rainfall map. In that way some of the stations were added or excluded from the catchment rainfall calculation despite they were previously considered or not in the Thiessen weighting. It was found out that the south and the western stations have the most significant influence on the catchment rainfall. There are seven sub-catchments (2 in the southern corner of Bhima Basin and 5 in the southern corner of Krishna basin) not having been identified with any rainfall station and G-D station, they are not considered in the initial rainfall modelling. However using parameter transfer process, they will ultimately be used in the overall basin models.

30 Interim Report 30 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6-2 Assigning rainfall station weitage for NAM catchment

Potential Evapotranspiration Potential Evapotranspiration (Ep) is input to the model. Although important it has less impact on the generated discharges than precipitation. Normally monthly average Ep data suffice and have been applied here. Ep input to the model has been based on Penman-Monteith estimates of an average year in various locations of the region. Penman-Monteith estimates are generally accepted as the most reliable Ep formula and normally deemed more reliable that pan evaporation measurements. At first, 13 stations with a complete record between 2000 and 2009 were chosen to be included in the NAM model. While visualising them on the map it was realised that there are some huge gaps. Therefore, additional 3 stations with 9 year record and were chosen and the missing data were reconstructed using correlation with the neighbouring stations. Thus, 16 evaporation stations were used in the NAM model and they were assigned to the catchments after visualising them on the terrain together with the average rainfall and potential evaporation values. Table 6.2 shows the stations from which evaporation data has been used in the model.

Interim Report 31 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Table 6.2 List of Evaporation Stations used in the NAM Rainfall-Runoff Model

Evaporation Station 1 2 3 4 5 6 7 8 Ambavade2 Barur Belwade Chaskman Dhulgaon Kashti Mhaisal Parali Name of NAM Catchment using the Evaporation Station Agrani R1 Bhima R12 Kera R1 Andhra R1 Agrani R2 Bhima R7 Krishna R10 Koyna R1 Yerala R1 Bori R2 Koyna R2 Andhra R2 Ghod R4 Krishna R11 Krishna R1 Yerala R2 Bori R2 Koyna R3 Bhama R1 Sina R1 Krishna R12 Krishna R2 Krishna R5 Bhama R2 Krishna R9 Krishna R3 Krishna R6 Bhima R1 Panchaganga R3Krishna R4 Krishna R7 Bhima R2 Mahu R1 Man R-K Bhima R3 Tarali R1 Morna R1 Bhima R4 Tarali R2 Morna R2 Ghod R1 Urmodi R1 Wang R1 Ghod R2 Urmodi R2 Wang R2 Indrayani R1 Vasana R1 Andhra R1 Indrayani R2 Veena R1 Indrayani R3 Indrayani R4 Indrayani R5 Kukadi R1 Kukadi R2 Mina R1 Pushpavati R1 Pushpavati R2 Pushpavati R3

32 Interim Report 32 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Evaporation Stations 9 10 11 12 13 14 15 16 Pargaon Patryachiwadi Sakhar Shigaon Wadange Late Rosa SidhewadiSol Name of NAM Catchment using the EvaporationStation Bhima R5 Bhogvati R1 Ambi R1 Krishna R8 Panchaganga R1 Nira R5 Bhima R8 Bhima R10 Bhima R6 Bhogvati R2 Gunjavani R1 Warana R2 Panchaganga R2 Nira R6 Bhima R9 Bhima R11 Ghod R3 Dhamni R1 Kanand R1 Warana R3 Nira R7 Sina R2 Man R-B Mula Mutha R1 Dhamni R2 Kanand R2 Sina R3 Sina R4 Mula Mutha R2 Dudhganga R1 Karha R1 Dudhganga R2 Mose R1 Kadavi R1 Mula R1 Kadavi R2 Mula R2 Kasari R1 Mula R3 Kasari R2 Mutha R1 Kumbhi R1 Mutha R2 Kumbhi R2 Mutha R3 Tulshi R1 Mutha R4 Nira R1 Nira R2 Nira R3 Nira R4 Pawana R1 Pawana R2 Pawana R3

Discharge Data Historical discharge data are required to calibrate the rain-fall runoff model, i.e. the simulated discharge are compared with the observed data for corresponding catchments. Two type of discharge time series were used in the Krishna-Bhima rainfall-runoff model. 1. Discharge from G-D stations: Historical time series of discharge from G-D stations as available were assigned to the appropriate catchments. Primarily the hourly values were chosen if they were available. However, during the calibration process it was realised that in case of running the model with daily rainfall values better results are achieved if considering daily discharge values. 2. Computed reservoir inflow series: The computations were checked for consistency, and in a few cases it was found that negative inflow values were computed. This could be either for accounting evaporation and other losses from the reservoirs or (for large negative values) due to inaccurate measurements of the reservoir water levels.

6.1.3 Model Parameters

Table 6.3 shows the calibrated parameters for each of the catchments.

Interim Report 33 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Table 6.3 NAM Parameters used in the Initial Calibration of Rainfall-Runoff Model

No. Catchment Name Umax Lmax CQOF CKIF CK1,2 TOF TIF TG CKBF Carea Sy GWLBF0 GWLBF1 Cqlow Cklow 1 PUSHPAVATI_R1 8 280 0.8 300 15 0.4 0.2 0.5 800 1 0.1 10 0 50 10000 2 PUSHPAVATI_R2 6 180 0.8 300 10 0.3 0.2 0.6 800 1 0.1 10 0 50 10000 3 PUSHPAVATI_R3 6 160 0.8 300 8 0.3 0.2 0.6 800 1 0.1 10 0 50 10000 4 GHOD_R1 12 220 0.75 500 25 0.45 0.4 0.6 2000 1 0.1 10 0 20 20000 5 BHIMA_R1 18 220 0.9 500 8 0.45 0.4 0.9 3000 1 0.1 10 0 10 30000 6 BHIMA_R2 18 220 0.9 400 8 0.35 0.4 0.75 1500 1 0.1 10 0 10 20000 7 KUKADI_R1 6 160 0.8 300 8 0.3 0.2 0.6 800 1 0.1 10 0 50 10000 8 ANDHRA_R1 14 300 0.9 400 30 0.65 0.7 0.8 3300 1 0.1 10 0 95 28000 9 INDRAYANI_R1 15 290 0.9 500 30 0.65 0.5 0.85 2500 1 0.1 10 0 65 16000 10 BHAMA_R1 18 220 0.9 500 8 0.45 0.4 0.9 3000 1 0.1 10 0 10 30000 11 ANDHRA_R2 10 300 0.9 600 24 0.6 0.7 0.8 2000 1 0.1 10 0 85 25000 12 INDRAYANI_R2 15 290 0.9 500 30 0.65 0.5 0.85 2500 1 0.1 10 0 65 16000 13 PAWANA_R1 10 280 0.88 500 17 0.65 0.5 0.9 2500 1 0.1 10 0 10 30000 14 MULA_R1 10 100 0.5 1000 10 0 0 0.7 1000 1 0.1 10 0 30 10000 15 MUTHA_R1 10 250 0.93 700 50 0.7 0.5 0.8 1500 1 0.1 10 0 10 26000 16 MOSE_R1 12 300 0.9 300 40 0.75 0.4 0.7 1000 1 0.1 10 0 30 15000 17 AMBI_R1 22 300 0.91 300 40 0.75 0.8 0.9 3000 1 0.1 10 0 30 15000 18 GHOD_R2 25 300 0.75 400 25 0.45 0.4 0.7 1500 1 0.1 10 0 20 20000 19 KUKADI_R2 25 300 0.75 400 25 0.45 0.4 0.7 1500 1 0.1 10 0 20 20000 20 GHOD_R3 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 21 GHOD_R4 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 22 KANAND_R1 16 280 0.85 700 35 0.85 0.5 0.65 1600 1 0.1 10 0 5 15000 23 NIRA_R1 22 280 0.91 500 35 0.6 0.7 0.8 2000 1 0.1 10 0 30 15000 24 NIRA_R2 13 240 0.82 450 35 0.45 0.7 0.85 2000 1 0.1 10 0 50 20000 25 NIRA_R3 22 280 0.85 700 35 0.6 0.7 0.8 3000 1 0.1 10 0 30 15000 26 KRISHNA_R1 24 190 0.95 300 23 0.8 0.4 0.7 1500 1 0.1 10 0 10 15000 34 Interim Report 34 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

27 KRISHNA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 28 MAHU_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 29 KOYNA_R1 18 320 0.72 600 28 0.55 0.5 0.75 1300 1 0.1 10 0 30 20000 30 VEENA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 31 URMODI_R1 14 250 0.95 500 22 0.9 0.5 0.8 2000 1 0.1 10 0 30 15000 32 TARALI_R1 14 230 0.91 300 17 0.7 0.5 0.68 1500 1 0.1 10 0 30 15000 33 MORNA_R1 20 170 0.8 300 15 0.65 0.3 0.75 1500 1 0.1 10 0 10 15000 34 WARANA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 35 KADAVI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 36 KASARI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 37 KUMBHI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 38 DHAMNI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 39 TULSHI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 40 DUDHGANGA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 41 BHOGVATI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 42 SINA_R1 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 43 SINA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 44 PAWANA_R2 12 280 0.88 300 8 0.65 0.8 0.9 2000 1 0.1 10 0 10 30000 45 INDRAYANI_R3 15 290 0.9 500 30 0.65 0.5 0.85 2500 1 0.1 10 0 65 16000 46 INDRAYANI_R4 15 290 0.9 500 30 0.65 0.5 0.85 2500 1 0.1 10 0 65 16000 47 INDRAYANI_R5 22 350 0.72 500 35 0.7 0.5 0.7 3000 1 0.1 10 0 20 30000 48 MUTHA_R2 18 300 0.88 400 30 0.6 0.7 0.9 2000 1 0.1 10 0 10 26000 49 MULA_R2 12 210 0.95 500 30 0.6 0.7 0.9 1000 1 0.1 10 0 20 20000 50 MUTHA_R3 20 300 0.85 700 60 0.6 0.4 0.7 3000 1 0.1 10 0 85 25000 51 TARALI_R2 16 250 0.88 300 10 0.65 0.4 0.72 1500 1 0.1 10 0 10 15000 52 MAN_R-K 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 53 WANG_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 54 MUTHA_R4 20 300 0.8 700 40 0.65 0.7 0.7 3000 1 0.1 10 0 85 25000

Interim Report 35 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

55 PAWANA_R3 14 280 0.74 300 17 0.65 0.8 0.9 2000 1 0.1 10 0 10 30000 56 BHIMA_R3 18 220 0.9 400 8 0.35 0.4 0.75 1500 1 0.1 10 0 10 20000 57 BHAMA_R2 18 220 0.9 400 8 0.35 0.4 0.75 1500 1 0.1 10 0 10 20000 58 MULA_R3 20 300 0.8 700 40 0.65 0.7 0.7 3000 1 0.1 10 0 85 25000 59 MULAMUTHA_R1 20 300 0.8 700 40 0.65 0.7 0.7 3000 1 0.1 10 0 85 25000 60 BHIMA_R6 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 61 MULAMUTHA_R2 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 62 BHIMA_R7 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 63 BHIMA_R8 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 64 GUNJAVANI_R1 18 240 0.93 500 9 0.55 0.7 0.8 2000 1 0.1 10 0 50 25000 65 KARHA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 66 KANAND_R2 14 300 0.82 900 20 0.5 0.4 0.8 3000 1 0.1 10 0 85 25000 67 NIRA_R4 20 300 0.82 900 60 0.6 0.7 0.7 3000 1 0.1 10 0 85 25000 68 KARHA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 69 NIRA_R5 20 300 0.82 900 60 0.6 0.7 0.7 3000 1 0.1 10 0 85 25000 70 BHIMA_R9 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 71 NIRA_R6 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 72 BHIMA_R10 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 73 KRISHNA_R3 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 74 VASANA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 75 URMODI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 76 KERA_R1 12 180 0.9 300 10 0.6 0.5 0.9 1500 1 0.1 10 0 10 15000 77 MORNA_R2 20 180 0.8 300 15 0.65 0.3 0.75 1500 1 0.1 10 0 10 15000 78 WANG_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 79 KADAVI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 80 YERALA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 81 WARANA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 82 KOYNA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000

36 Interim Report 36 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

83 KRISHNA_R4 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 84 KRISHNA_R5 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 85 KRISHNA_R6 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 86 KRISHNA_R7 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 87 YERALA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 88 KRISHNA_R8 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 89 KRISHNA_R9 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 90 KRISHNA_R10 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 91 KASARI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 92 KUMBHI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 93 DHAMNI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 94 BHOGVATI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 95 PANCHAGANGA_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 96 PANCHAGANGA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 97 KRISHNA_R12 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 98 DUDHGANGA_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 99 AGRANI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 100 AGRANI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 101 MAN_R-B 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 102 BHIMA_R11 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 103 SINA_R4 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 104 BHIMA_R12 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 105 MINA_R1 12 300 0.75 500 22 0.35 0.4 0.75 2000 1 0.1 10 0 10 20000 106 BHIMA_R4 25 350 0.72 500 30 0.7 0.4 0.75 3000 1 0.1 10 0 20 30000 107 BHIMA_R5 18 220 0.9 400 8 0.35 0.4 0.75 1500 1 0.1 10 0 10 20000 108 PANCHAGANGA_R3 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 109 KRISHNA_R11 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 110 WARANA_R3 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000

Interim Report 37 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

111 BORI_R1 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 112 KOYNA_R3 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 113 NIRA_R7 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 114 BORI_R2 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000 115 SINA_R3 10 100 0.5 1000 10 0 0 0 2000 1 0.1 10 0 0 10000

38 Interim Report 38 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

6.2 Model Outputs The primary output of the hydrological model is inflow to the river models from the sub-catchments, giving a distributed inflow pattern to the river systems and reservoirs. The hydrological input to the river models allows short and long term forecasting based on real time updating, and accounting for the near hydrological history including the previous monsoon, and rainfall over the preceding days.

6.3 Calibration

Calibration Procedure The hydrological model has been calibrated by adjusting the model parameters and to a certain degree the unknown distribution of precipitation over the area to obtain the best possible model performance in terms of its ability to replicate the historical observed hydrographs on the basis of the historical input. If the model is able to simulate the historical hydrographs well it will also perform well in future simulations. In the case of the NAM model this concept has been tested through numerous projects, scientific studies as well as operational forecasting systems. After it was ascertained that the derived catchment rainfall was more or less following the water balance the calibration of the parameters was initiated. Here the auto calibration module was used only as a guidance what could be possibly improved but only if the parameters seemed realistic. The final model parameters were chosen so the best compromise was achieved between these three criteria: matching the peaks, matching the cumulative water balance (Wbl) curve and reaching as high as possible the coefficient of determination (R2). The water balance error (Wbl) is attempted to be as low as possible.

Calibration Period The calibration period is from 2002 to 2009.

Calibration Results The model has been calibrated on all the available discharge gauging stations and reservoir inflow data. The calibration results for a simulation without data assimilation (correcting the model with real time data) are shown for a number of strategic locations. Sample calibration plots are presented in Figures 6.3 to 6.10. In the calibration plots R2 denotes the fitness between simulated and observed flows, while Wbl means the water balance error in mm/year. The calibration exercises will continue to improve the fitness between the observed and simulated values. In all cases, the opportunity for a much improved calibration will arise with the availability of the higher frequency and higher density observations from the RTDAS. In addition, the advanced data assimilation in the MIKE 11 software, compensates for any deficiencies in the weather forecast and real time data, and in the model setup, in relation to the actual current catchment response.

Interim Report 39 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6-3: Comparison of Simulated and Observed Discharges for Catchment Bhima_R2 (R2=0.80, Wbl=-1.2% (Obs=1886mm/y, Sim=1863mm/y))

40 Interim Report 40 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6-4: Simulated and Observed Inflows to Koyna Reservoir (R2=0.87, Wbl=3.60% (Obs=5517mm/y, Sim=5311mm/y))

Interim Report 41 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6-5: Comparison of Simulated and Observed Flows for Catchment Gunjavani_R1 (R2=0.82, Wbl=-4.0% (Obs=2552mm/y, Sim=2654mm/y))

42 Interim Report 42 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6.6: Comparison of Simulated and Observed Flows for Catchment Urmodi_R1 (R2=0.81, Wbl=-1.70% (Obs=2204mm/y, Sim=2240mm/y))

Interim Report 43 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6.7 Comparison of Simulated and Observed Flows for Catchment Morna_R2 (R2=0.83, Wbl=-1.70% (Obs=3246mm/y, Sim=3170mm/y))

44 Interim Report 44 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6.8: Comparison of Simulated and Observed Flows for Catchment Mutha_R2 (R2=0.81, Wbl=-3.51% (Obs=2374mm/y, Sim=2456mm/y))

Interim Report 45 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 6.9: Comparison of Simulated and Observed Flows for Catchment Pawana_R1 (R2=0.67, Wbl=6.10% (Obs=3565mm/y, Sim=3346mm/y))

46 Interim Report 46 Krishna & Bhima River Basins RTSF&ROS RTSF&ROS

Figure 610: Comparison of Simulated and Observed Flows for Catchment Mose_R1 (R2=0.74, Wbl=1.60% (Obs=3864mm/y, Sim=3803mm/y))

Interim Report 47 RTSF&ROS Krishna & Bhima River Basins

6.4 Further Work

The NAM Rainfall-Runoff model of the Krishna and Bhima River basins is the key model, which simulates runoff from the catchment for a given rainfall time series. Therefore, it has to be calibrated to the highest accuracy possible. In the Krishna and Bhima basins runoff from most of the important catchments are inflows to the reservoirs. Only in a few cases the G-D stations capture the runoff s without any storage effect of the reservoirs. Therefore, the final calibration of the NAM model should be done in conjunction with the MIKE11 hydrodynamic model, which incorporates dynamic routing of flows through the river and reservoir system. In the initial calibration of the model both daily and hourly rainfall data have been used. However, in the real time forecasting application, hourly data will be sued. Hence, the model needs to be fine-tuned with hourly data when the RTDAS is in place. Simulation of runoffs from the seven ungagged catchments will be carried out using data from nearby catchments.

48 Interim Report Krishna & Bhima River Basins RTSF&ROS

7 RIVER BASIN SIMULATION MODEL 7.1 Model Setup

7.1.1 Network The river network and catchments have been delineated from the 90m Digital Elevation Model from NASA, using the semi-automatic routine in MIKE BASIN. The river network and corresponding catchments are illustrated in Figure 7-1. Figure 7-2 shows the detailed network of the Koyna-Krishna part of the basin, which illustrates that details of any part of the basin can be viewed. The sub- catchments in the river basin model correspond to those in the hydrological model. The discharge from each sub-catchment enters the river basin model at the most downstream point of each catchment.

Figure 7-1 Krishna-Bhima MIKE BASIN Network

Interim Report 49 RTSF&ROS Krishna & Bhima River Basins

Figure 7-2 MIKE BASIN Network details for part of Koyna-Krishna area 7.1.2 Infrastructure The reservoirs, diversions canals and hydropower plants included in the river basin model are also illustrated in Figure 7-1. Forty six (46) reservoirs in the two basins are included in the present version of the model. Reservoir Characteristics The main characteristics of the reservoirs are implemented in the form of level- storage-surface characteristics, maximum and minimum operation levels, mandatory minimum releases to the downstream river, and flood control rule curves. The latter curves are for all reservoirs set at a constant level, above which the model will start spilling. In some hydropower plants, spilling only takes place when the total turbine capacity is exceeded, i.e. it is assumed that as much excess water as possible is utilised for power production.

Hydropower Characteristics The characteristics of the hydropower stations are included in the form of their maximum capacity, tail water levels (head water levels are given by the reservoir) and an efficiency factor of 80 %.

7.1.3 Initial Conditions For long term simulation, the most important initial condition for the model is the correct storage in all reservoirs. These volumes will be calculated by the model from the latest available observations of reservoir water levels received from the RTDAS.

7.1.4 Model Inputs The Main inputs to the model are:

50 Interim Report Krishna & Bhima River Basins RTSF&ROS

 Inflow to the main rivers in the form of runoff from the sub-catchments. The runoff will be available for the entire simulation period.  Release requirements from the reservoirs. The model is open for two sources of input. It can use the gauged inflow series, which are used directly for the gauged catchments or transferred in the form of area specific runoff (in l/s/km2) to nearby un-gauged catchments. This input gives the best simulation of the historical period but is not the best option for future predictions. Alternatively the model can use the discharges generated by the hydrological model, which does not give the best representation of the past but, owing to the embedded memory of near past events in the hydrological model and the planned use of explicit or implicit meteorological forecasts, this is the best method for future predictions. The release requirements from the main reservoirs must be entered into the model.

7.1.5 Irrigation Data The data collected from management divisions of WRD on irrigation planning, areas under different crops, information on soils and climatological data collected from HP have been used as input to the Irrigation Module of Mike Basin, which is based on FAO 56 methodology. Figure 7-3 a shows two examples of providing irrigation water demand, monthly and daily basis.

Figure 7-3 Example of Irrigation Water Demand (monthly above, daily below)

7.1.6 Water Supply Data The water demand for non-irrigation users like domestic water use (drinking & other) and Industrial has been collected from WRD. The time series of these demands are included in the models on monthly basis as the seasonal water demands are not as dynamic as irrigation.

Interim Report 51 RTSF&ROS Krishna & Bhima River Basins 7.2 Model Outputs The model simulates the flows in all represented rivers and canals on a daily basis during the simulation period. The most important output from the model will be the simulated inflows to and the storage in the reservoirs, and the reservoir releases and power generation being a consequence of the selection of wet and dry future inflows, or according to long term weather forecasts. The model seeks to fulfil the requested releases from the reservoirs, and keeps a record on the deficits in these releases owing to constrained inflows or storage. Owing to the difficulty in predicting the meteorology months into the future, the model results are normally presented in the form of probabilistic time series that will be derived from a Monte-Carlo simulation of the coming week/month/year, where each of the one yearly simulations in the sequence is initiated with the latest real time remote sensing and RTDAS information, followed by the input of meteorological data from a selected individual historical year. Considering that the routing time in the river system (hours) is much smaller than the long term forecast period (weeks, months), no flow routing is included in the river basin simulation model. The established model is a pure water balance model and the calibration exercise is one of checking the consistency of the introduced characteristics, mainly with respect to the reservoirs. Sample outputs of the MIKE BASIN Model are presented in Figures 7-4 to 7-7. Figure 7-4 shows simulated water level and storage for Ujjani Reservoir. Figure 7- 5 shows release from the reservoir. Figure 7.6 shows a comparison of irrigation demand and supply. Figure 7-7 shows an example of water supply and demand comparison.

Figure 7-4 Simulated Water Level and Storage of Ujjani Reservoir

52 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 7-5 Simulated Release from Ujjani Reservoir

Figure 7-6 Example of Irrigation Water supply and deficit

Figure 7-7 Example of water supply and deficit

7.3 Further Work

Once all required data from all reservoirs are available and the NAM Rainfall- Runoff model is finally calibrated, the MIKEBASIN model will be finalised for

Interim Report 53 RTSF&ROS Krishna & Bhima River Basins

simulating historical conditions and compared with observed/computed values. As discussed above, the model will be used in long term-forecasting of reservoir release based on the present status available from the RTDAS and with assumed probabilistic forecasts of parameters like rainfall and water demand on weekly or monthly basis. The MIKEBASIN model will also be used in optimising reservoir operation for a variety of objective functions and constraints for both single objective as well as multi-objective optimization.

54 Interim Report Krishna & Bhima River Basins RTSF&ROS

8 HYDRODYNAMIC MODEL 8.1 Introduction The Hydrodynamic River Model takes the rainfall-runoff from the NAM, and carries out a continuous routing of the flows and flood waves through the main rivers and reservoirs of the basin. The model outputs discharges and water levels throughout, for application to short term Flood Forecasting and Optimisation. 8.2 Model Setup

8.2.1 Network The hydrodynamic river model for short term flood forecasting is established for the two basins combined (Figure 8-1). A detailed network zoomed for an area is shown in Figure 8-2. The model describes the propagation of flood waves through the river and reservoir system, incorporating data assimilation at all the real time discharge and water level stations in the system.

Figure 8-1: MIKE 11 Model Network showing runoff catchments

Interim Report 55 RTSF&ROS Krishna & Bhima River Basins

Figure 8-2 Details of the MIKE11 network zoomed to an area The computational network presented in Figure 8-1and Error! Reference source not found. is established applying all available information. I.e. river and reservoir shape files and geo-referred satellite images provided by WRD plus information extracted from Google Earth.

Cross Sections Since only a limited number of river cross sections area available, the pilot MIKE 11 model is constructed using interpolated cross sections. The interpolation process is tedious as an interpolated cross section should represent the conveyance capacity of the river channel in proportion to the catchment area upstream interpolation point. Also the river bottom level should follow the natural slope. The SRTM generated DEM was used to adjust the river bed levels accordingly. Table 8.1 shows the details of river cross sections. A large number of cross sections were available for the from Khadakwasala dam to a length of 27 km surveyed at an interval of 30m. However, since this close interval of cross sections is not required for river modelling, cross sections at every 300m was extracted and used in a separate pilot model of the Mutha River.

56 Interim Report Krishna & Bhima River Basins RTSF&ROS

Table 8.1 Detailed description of river branches

Chainage (km) Cross Sections Length Sr. No River Branch (km) Nos. Nos. U/S D/S Source available interpolated Bhima Basin 1 Ambi 0.000 2.012 2.012 0 NA 2 2 Andhra-DS 0.000 2.833 2.833 0 NA 2 3 Bhama 0.000 34.470 34.470 0 WRD 8 4 Bhima 0.000 36.545 36.545 1 WRD 7 5 Bhima 36.550 50.405 13.855 0 NA 4 6 Bhima 50.415 109.840 59.425 1 CWC 13 7 Bhima 109.860 140.605 30.745 0 NA 8 8 Bhima 140.615 312.260 171.645 2 CWC 37 9 Bhima 312.265 418.265 106.000 1 WRD 22 10 Bhima 418.275 502.675 84.400 1 CWC 18 11 Bhima 502.680 526.958 24.278 1 WRD 5 12 Bhogawati-Sina 0.000 7.330 7.330 1 WRD 2 13 Chilhewadi-DS 0.000 18.600 18.600 0 NA 5 14 Ghod 0.000 48.245 48.245 0 NA 11 15 Ghod 48.265 90.800 42.535 0 NA 10 16 Ghod 90.810 159.012 68.202 0 NA 16 17 Gunjawani 0.000 4.255 4.255 1 WRD 1 18 Indrayani 0.000 17.750 17.750 0 NA 5 19 Indrayani 17.760 103.188 85.428 1 WRD 18 20 Kanand 0.000 42.140 42.140 2 WRD 8 21 Karha 0.000 68.905 68.905 1 WRD 15

Interim Report 57 RTSF&ROS Krishna & Bhima River Basins

Chainage (km) Cross Sections Length Sr. No River Branch (km) Nos. Nos. U/S D/S Source available interpolated 22 Kukadi 0.000 28.690 28.690 0 NA 7 23 Kukadi 28.700 93.546 64.846 0 NA 15 24 Kusegaon-DS 0.000 9.128 9.128 0 NA 3 25 Man-Bhima 0.000 160.825 160.825 1 WRD 33 26 Mina 0.000 40.076 40.076 0 NA 10 27 Mose 0.000 2.238 2.238 0 NA 2 28 Mose 2.245 8.688 6.443 0 NA 2 29 Mula 0.000 43.795 43.795 1 WRD 10 30 Mula 43.800 55.787 11.987 0 NA 4 31 Mula-Mutha 55.792 125.907 70.115 1 WRD 15 32 Mutha 0.000 19.322 19.322 1 WRD 5 33 Mutha 19.332 47.706 28.374 1 WRD 7 34 Nira 0.000 34.975 34.975 1 WRD 9 35 Nira 34.985 131.500 96.515 1 WRD 20 36 Nira 131.510 216.565 85.055 1 WRD 18 37 Pawana 0.000 57.714 57.714 1 WRD 13 38 Pushpawati 0.000 14.107 14.107 0 NA 4 39 Pushpawati 14.127 24.326 10.199 0 NA 3 40 Shirota-DS 0.000 5.130 5.130 0 NA 2 CWC, 41 0.000 335.545 335.545 3 68 Sina WRD 42 Wadiwale-DS 0.000 8.194 8.194 1 WRD 3 Krishna Basin 1 Agrani 0.000 5.320 5.320 1 WRD 1

58 Interim Report Krishna & Bhima River Basins RTSF&ROS

Chainage (km) Cross Sections Length Sr. No River Branch (km) Nos. Nos. U/S D/S Source available interpolated 2 Bhogavati 0.000 59.560 59.560 0 NA 13 3 Bhogavati 59.570 63.560 3.990 0 NA 2 4 Bhogavati 63.570 76.545 12.975 0 NA 5 5 Chikutra 0.000 7.895 7.895 1 WRD 2 6 Dhamani 0.000 32.625 32.625 1 WRD 8 7 Dudhganga 0.000 69.960 69.960 2 WRD 14 8 Ghataprabha 0.000 2.710 2.710 1 WRD 1 9 Hiranyakeshi 0.000 55.225 55.225 2 WRD 11 10 Kadavi 0.000 40.165 40.165 1 WRD 8 11 Kasari 0.000 73.050 73.050 1 WRD 15 12 Kera 0.000 2.435 2.435 1 WRD 1 13 Koyna 0.000 26.290 26.290 1 CWC 6 14 Koyna 26.300 33.815 7.515 0 NA 3 15 Koyna 33.825 67.820 33.995 1 CWC 8 16 Krishna 0.000 53.695 53.695 0 NA 12 17 Krishna 53.705 66.807 13.102 0 NA 4 18 Krishna 66.820 95.920 29.100 0 NA 7 19 Krishna 95.930 132.980 37.050 1 WRD 9 CWC, 20 132.990 230.665 97.675 2 18 Krishna WRD 21 Krishna 230.675 245.310 14.635 2 WRD 2 CWC, 22 245.320 283.055 37.735 4 2 Krishna WRD 23 Krishna 283.065 293.029 9.964 3 WRD 1 24 Kumbhi 0.000 36.485 36.485 1 WRD 9

Interim Report 59 RTSF&ROS Krishna & Bhima River Basins

Chainage (km) Cross Sections Length Sr. No River Branch (km) Nos. Nos. U/S D/S Source available interpolated 25 Kumbhi 36.495 52.425 15.930 0 NA 5 26 Mahu DS 0.000 23.217 23.217 0 NA 6 27 Morna 0.000 18.245 18.245 1 WRD 4 28 Panchaganga 0.000 79.900 79.900 5 WRD 5 29 Panchaganga 79.920 80.756 0.836 1 WRD 1 30 PanchagangaLoop 0.000 2.190 2.190 0 NA 2 31 Tamraparni 0.000 64.310 64.310 2 WRD 11 32 Tarali 0.000 31.365 31.365 1 WRD 7 33 Tulshi 0.000 30.945 30.945 0 NA 8 34 Urmodi 0.000 36.450 36.450 1 WRD 9 35 Uttarmand DS 0.000 16.435 16.435 0 NA 5 36 Vasana 0.000 4.642 4.642 1 WRD 1 37 Vedganga 0.000 10.875 10.875 1 WRD 3 38 Vedganga 10.880 14.540 3.660 1 CWC 1 39 Veena 0.000 18.678 18.678 0 NA 5 40 Wang 0.000 24.320 24.320 1 WRD 5 41 Warana 0.000 50.900 50.900 0 NA 11 42 Warana 50.910 129.535 78.625 2 WRD 15 43 WaranaLoop 0.000 1.245 1.245 0 NA 2 44 Yerala 0.000 82.715 82.715 1 WRD 17

60 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figures 8-3 and 8-4 show examples of surveyed and interpolated cross sections.

Figure 8-3: Example of River Cross Sections surveyed and interpolated

The above model network was used for the pilot model. More detailed cross sections will be required for developing an accurate hydrodynamic model to be useful for discharge and water level forecasting. It is expected that new cross sections will available in about 2 to 3 months time after the river survey programme is implemented by WRD. Figures 8-4 and 8-5 show some examples of longitudinal sections of the rivers based on available and interpolated cross sections.

Interim Report 61 RTSF&ROS Krishna & Bhima River Basins

Figure 8-4 Longitudinal Section (L-S) along Koyna-Krishna Rivers

62 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 8-5 Longitudinal Section (L-S) along Ghod-Bhima Rivers

Interim Report 63 RTSF&ROS Krishna & Bhima River Basins

Reservoirs Elevation:area:volume (EAV) curves have been obtained for the reservoirs based on the topographic data at the time of dam design. MIKE 11 calculates as default the surface area at a specific grid point as the level-dependent cross-section width times half the distance from its upstream to its downstream cross-section. Reservoir cross-sections primarily represent the correct flow conveyance of the individual transects; not necessarily the correct stage - storage relation. To adjust the MIKE 11 model to comply with the EAV relations supplied by WRD a number of local sub-models representing individual reservoir were developed. Synchronous levels and surface areas were extracted from these models and the respective “Additional Storage Area” parameters in the MIKE 11 cross-section database were adjusted accordingly until a match between the MIKE 11 and the WRD EAV relations were obtained. As an illustration of this process the schematisation of the Ujjani reservoir is described below (Figure 8-6).

Figure 8-6 Ujjani river network and reservoir cross sections In the above figure (8-6) the Ujjani reservoir is shown together with extracted reservoir cross-sections. These cross-sections are extracted from the SRTM DEM and detail are available only above the water surface (Figures 8-7 and 8-8).

Figure 8-7 Reservoir-cross section extracted from a SRTM DEM 64 Interim Report Krishna & Bhima River Basins RTSF&ROS

Based on up- and downstream river bed levels and bank slopes, actual reservoir cross-sections were interpolated. See below:

Figure 8-8 Reservoir cross-section adjusted below the water surface Based on these interpolated reservoir cross-sections the “RAW” “stage-storage” relation (EAV) was extracted from the MIKE 11 cross-section database and compared with the EAV relation supplied by WRD. Finally the MIKE 11 cross- sections were adjusted until a perfect match between the model and the nature was obtained. Raw and adjusted stage – surface area curves are shown below in Figure 8-9.

Figure 8-9 Reservoir EAV relations from MIKE 11 (red) and WRD Data ( blue)

Flood plain levels In order to produce flood inundation maps accurate flood plain transects and a high resolution digital elevation model (DEM) is required. Since neither of these area presently available, the flood plain modelling and hence flood mapping must be postponed until these topographical information become available. WRD has been advised to extend levelling surveys well into the flood plains and a DEM will be developed based on these flood plain levels.

Interim Report 65 RTSF&ROS Krishna & Bhima River Basins

Structures A variety of structures can be included in the hydrodynamic model. In the Krishna- Bhima model, reservoir spillway gates, irrigation outlets, hydro power releases and leakage are incorporated as “control structures”. The functions of the gated control structures can be simulated for different types of control variables, such as water level, discharge, gate level etc. Figure 8-10 shows an example of control structure definition in the model.

Figure 8-10 Example of control structure definition in MIKE11 model

Reservoir operation and the related configuration of Control Structures are further detailed in Section 8.4

8.2.2 Boundary conditions Catchment runoffs from the NAM Rainfall-Runoff model are used as upstream boundaries and intermediate inflows. The hydrodynamic model has an automatic coupling to the rainfall-runoff model. The entire area of the two basins is subdivided into 122 catchments. Each catchment is connected to the river model either by a point connection in the case of a major tributary, or distributed in the case of minor tributaries. Table 8.2 shows the details of catchment connections to the river network.

66 Interim Report Krishna & Bhima River Basins RTSF&ROS

Table 8.2 Connection of Runoff Catchments to River Network

Area U/S D/S (sq. Chainage Chainage Sr. No. Catchment Name km) River Name (km) (km) 1 SINA_R1 1474.7 Sina 0 751.29 2 SINA_R2 4154.3 Sina 751.29 1648.21 3 SINA_R3 2997.4 Sina 1648.21 2449.02 4 SINA_R4 3636.3 Sina 2449.02 3239.61 5 BHIMA_R1 45 Kalmodi Reservoir 0 0 6 BHIMA_R2 254.2 Chaskaman Reservoir 0 120 7 BHIMA_R3 85.7 Bhima 0 67 8 BHIMA_R4 300 Bhima 67 365.45 9 BHIMA_R4 50 Bhima 365.5 504.05 10 BHIMA_R4 150 Bhama 67 344.7 11 BHIMA_R4 235.5 Indrayani 652.45 1031.88 12 BHIMA_R5 66.2 Bhima 504.15 574 13 BHIMA_R6 1005 Bhima 574 1098.4 14 BHIMA_R7 630 Bhima 1098.6 1406.05 15 BHIMA_R7 48.2 Bhima 1406.15 1450.03 16 BHIMA_R8 3620.7 Bhima 1450.03 2740 17 BHIMA_R9 386.3 Bhima 2760 3122.6 18 BHIMA_R9 100 Nira 1884.81 2155.65 19 BHIMA_R10 1929.6 Bhima 3122.65 3804.35 20 BHIMA_R11 700 Bhima 3804.35 4182.65 21 BHIMA_R11 50 Man-Bhima 1470.84 1608.25 22 BHIMA_R11 3842.6 Bhima 4182.75 4892.55 23 BHIMA_R12 50 Bhima 4893 5026.75 24 BHIMA_R12 1159.4 Bhima 5026.8 5269.58 25 BHIMA_R12 50 Sina 3245.64 3355.45 26 GHOD_R1 273.7 Dimbe Reservoir 0 0 27 GHOD_R2 500 Ghod 0 482.45 28 GHOD_R2 200 Mina 0 400.76 29 GHOD_R2 371.1 Ghod 482.65 836 30 GHOD_R3 546.7 Ghod 952.4 1260 31 GHOD_R4 912.3 Ghod 1290 1516 32 MINA_R1 133.7 Wadaj Reservoir 0 0 33 PUSHPAVATI_R1 104 Chilewadi Reservoir 0 0 Pimplegaon Joga 34 PUSHPAVATI_R2 90.5 Reservoir 0 0 35 PUSHPAVATI_R3 106.2 Manikdoh Reservoir 0 0 36 KUKADI_R1 180 Kukadi 0 280 37 KUKADI_R1 80 Pushpawati 0 141.07 38 KUKADI_R1 20 Pushpawati 141.27 243.26 39 KUKADI_R1 89.3 Chilhewadi-DS 0 186 Interim Report 67 RTSF&ROS Krishna & Bhima River Basins

Area U/S D/S (sq. Chainage Chainage Sr. No. Catchment Name km) River Name (km) (km) 40 KUKADI_R2 923.1 Kukadi 287 935.46 Bhama Aksheda 41 BHAMA_R1 191.9 Reservoir 0 0 42 BHAMA_R2 37.8 Bhama 0 62 43 MULAMUTHA_R1 638.8 Mula-Mutha 600 1060 44 MULAMUTHA_R2 243.3 Mula-Mutha 1060 1259.07 45 MULA_R1 248.9 Mulshi Reservoir 0 0 46 MULA_R2 85.9 Mula 0 128 47 MULA_R3 450 Mula 130 437.65 48 MULA_R3 148.1 Mula 438 557.87 49 PAWANA_R1 116.6 Pawana Reservoir 0 0 50 PAWANA_R2 39 Kasar Sai Reservoir 0 0 51 PAWANA_R3 270.1 Pawana 0 577.14 52 PAWANA_R3 40 Kusegaon-DS 0 91.28 53 MUTHA_R1 36 Temghar Reservoir 0 0 54 MOSE_R1 131.6 Warasgaon Reservoir 0 0 55 AMBI_R1 118.2 Panshet Reservoir 0 0 56 MUTHA_R2 46.8 Mutha 0 82 57 MUTHA_R3 30 Mose 22.45 86.88 58 MUTHA_R3 50 Mutha 0 193 59 MUTHA_R3 145.1 Mutha 193.32 320 60 MUTHA_R4 124.4 Mutha 330 440 61 INDRAYANI_R2 14.7 Valvan Reservoir 0 0 62 INDRAYANI_R3 30.6 Shirota Reservoir 0 0 63 INDRAYANI_R1 45.6 Wadiwale Reservoir 0 0 64 ANDHRA_R1 124.1 Nethersole Reservoir 0 0 65 ANDHRA_R2 90.7 Andhra Reservoir 0 160 66 INDRAYANI_R4 11 Wadiwale-DS 0 30 67 INDRAYANI_R5 40 Indrayani 0 177.5 68 INDRAYANI_R5 350.6 Indrayani 180 650 69 NIRA_R1 111 Nira Deoghar Reservoir 0 0 70 NIRA_R2 273.6 Bhatghar Reservoir 0 0 71 NIRA_R3 131.3 Nira 0 114 72 NIRA_R4 200 Nira 120 346.95 73 NIRA_R4 509.6 Nira 349.85 575 74 NIRA_R5 1243.2 Nira 580 974 75 NIRA_R6 200 Nira 980 1315 76 NIRA_R6 847.9 Karha 456 689.05 77 NIRA_R7 1585.7 Nira 1315.1 1860 78 KARHA_R1 397.1 Nazare Reservoir 0 0 79 KARHA_R2 734.8 Karha 0 450 80 KANAND_R1 51.5 Gunjawani Reservoir 0 0

68 Interim Report Krishna & Bhima River Basins RTSF&ROS

Area U/S D/S (sq. Chainage Chainage Sr. No. Catchment Name km) River Name (km) (km) 81 GUNJAVANI_R1 42 Kanand 141 141 82 KANAND_R2 432.1 Kanand 0 421.4 83 MAN_R-B 4614.5 Man-Bhima 0 1608.25 84 INDRAYANI_R5 20 Andhra-DS 0 28.33 85 INDRAYANI_R4 10 Shirota-DS 0 51.3 86 INDRAYANI_R5 30 Wadiwale-DS 0 81.94 Dhom Balkawadi 87 KRISHNA_R1 39.8 Reservoir 0 0 88 KRISHNA_R2 155.5 Dhom Reservoir 0 120 89 MAHU_R1 28.9 Mahu DS 0 0 90 VEENA_R1 212.3 Kanher Reservoir 0 0 91 KRISHNA_R3 100 Veena 0 186.78 92 KRISHNA_R3 100 Mahu DS 0 232.17 93 KRISHNA_R3 637 Krishna 0 536.95 94 URMODI_R1 112.7 Urmodi Reservoir 0 0 95 URMODI_R2 22.2 Urmodi 0 20 96 TARALI_R1 80.5 Tarali Reservoir 0 0 97 TARALI_R2 55.2 Tarali 0 60 98 MAN_R-K 39.9 Uttarmand DS 0 0 99 VASANA_R1 527 Vasana 0 0 100 KRISHNA_R4 25 Vasana 0 46.22 101 KRISHNA_R4 50 Uttarmand DS 0 164.35 102 KRISHNA_R4 300 Urmodi 0 364.5 103 KRISHNA_R4 150 Tarali 0 313.65 104 KRISHNA_R4 50 Krishna 537.05 668.07 105 KRISHNA_R4 425 Krishna 668.2 959.2 106 KRISHNA_R4 165.5 Krishna 959.3 1132 107 KRISHNA_R5 220 Krishna 1132 1329.8 108 KRISHNA_R5 19.2 Koyna 645 678.2 109 KRISHNA_R6 31.3 Krishna 1330 1400 110 KRISHNA_R7 710.5 Krishna 1400 1895 111 YERALA_R1 766.3 Yerala 0 0 112 YERALA_R2 1317.4 Yerala 0 355 113 KRISHNA_R8 1400 Yerala 360 827.15 114 KRISHNA_R8 354.1 Krishna 1897 2306.65 115 KRISHNA_R8 25 Krishna 2306.75 2400 116 KRISHNA_R9 50 Krishna 2400 2453.1 117 KRISHNA_R9 77.1 Krishna 2453.2 2510 118 KRISHNA_R9 50 Warana 1200 1295.35 119 KRISHNA_R10 440.3 Krishna 2460 2695 120 KRISHNA_R11 120.6 Krishna 2700 2830.55 121 KRISHNA_R12 21.8 Krishna 2830.65 2865

Interim Report 69 RTSF&ROS Krishna & Bhima River Basins

Area U/S D/S (sq. Chainage Chainage Sr. No. Catchment Name km) River Name (km) (km) 122 KRISHNA_R13 872.9 Krishna 2870 2930.29 123 KRISHNA_R13 100 Dudhganga 570 699.6 124 KOYNA_R1 897.9 Koyna Reservoir 0 600 125 KERA_R1 107.4 Kera 0 0 126 MORNA_R1 54.3 Morna 0 0 127 MORNA_R2 60.6 Morna 0 140 128 KOYNA_R2 200 Koyna 0 262.9 129 KOYNA_R2 40 Koyna 263 338.15 130 KOYNA_R2 11.6 Koyna 338.25 355 131 KOYNA_R2 10 Morna 145 182.25 132 WANG_R1 69.9 Wang 0 0 133 WANG_R2 162.9 Wang 0 75 134 KOYNA_R3 125 Wang 80 243.2 135 KOYNA_R3 180 Koyna 360 640 136 AGRANI_R1 139.2 Salpawadi Reservoir 0 0 137 AGRANI_R2 480.5 Agrani 0 370 138 AGRANI_R3 862 Agrani 370 420 139 WARANA_R1 274.1 Warana Reservoir 0 0 140 KADAVI_R1 24 Kadvi Reservoir 0 0 141 KADAVI_R2 330.3 Kadavi 0 350 142 WARANA_R2 300 Warana 0 509 143 WARANA_R2 851.4 Warana 509.1 930 144 WARANA_R3 248.9 Warana 935 1195 145 KASARI_R1 32.7 Kasari Reservoir 0 0 146 KASARI_R2 552.4 Kasari 0 660 147 PANCHAGANGA_R1 50 Kasari 665 730.5 148 PANCHAGANGA_R1 25 Bhogavati 700 765.45 149 PANCHAGANGA_R1 186.1 Panchaganga 0 230 150 PANCHAGANGA_R2 427.7 Panchaganga 235 510 151 PANCHAGANGA_R3 113.8 Panchaganga 515 650 152 KRISHNA_R11 10 PanchagangaLoop 0 21.9 153 KRISHNA_R11 40 Panchaganga 655 799 154 KUMBHI_R1 20.8 Kumbhi Reservoir 0 0 155 KUMBHI_R2 98.7 Kumbhi 0 190 156 BHOGVATI_R2 100 Kumbhi 195 364.85 157 DHAMNI_R1 42.6 Dhamani 0 0 158 DHAMNI_R2 111 Dhamani 0 180 159 BHOGVATI_R2 25 Dhamani 185 326.25 160 TULSHI_R1 34.3 Tulshi Reservoir 0 0 161 BHOGVATI_R2 130 Tulshi 0 309.45 162 BHOGVATI_R1 132.2 Radhanagari Reservoir 0 0 163 BHOGVATI_R2 200 Bhogavati 0 595.6

70 Interim Report Krishna & Bhima River Basins RTSF&ROS

Area U/S D/S (sq. Chainage Chainage Sr. No. Catchment Name km) River Name (km) (km) 164 BHOGVATI_R2 25 Bhogavati 595.7 635.6 165 BHOGVATI_R2 50 Bhogavati 635.7 695 166 BHOGVATI_R2 110.1 Kumbhi 364.95 524.25 167 DUDHGANGA_R1 203.5 Dudhganga Reservoir 0 0 168 DUDHGANGA_R2 400.2 Dudhganga 0 565 169 VEDGANGA_R1 600 Vedganga 0 0 170 VEDGANGA_R1 300 Chikutra 0 0 171 VEDGANGA_R1 60 Vedganga 0 108.75 172 VEDGANGA_R1 40.5 Vedganga 108.8 145.4 173 VEDGANGA_R1 60 Chikutra 0 78.95

Interim Report 71 RTSF&ROS Krishna & Bhima River Basins

Table 8.3 shows all the boundary conditions used in the MIKE11 model.

Table 8.3 Boundary Conditions used in the MIKE11 Model Sr. No. Boundary Type Name Chainage (km) 1 Inflow Panshet Reservoir 0 2 Inflow Nethersole Reservoir 0 3 Inflow Bhama Aksheda Reservoir 0 4 Q-h Bhima 526.96 5 Inflow Chilewadi Reservoir 0 6 Inflow Dimbe Reservoir 0 7 Inflow Gunjawani 0 8 Inflow Valvan Reservoir 0 9 Inflow Gunjawani Reservoir 0 10 Inflow Nazare Reservoir 0 11 Inflow Manikdoh Reservoir 0 12 Inflow Kasar Sai Reservoir 0 13 Inflow Man-Bhima 0 14 Inflow Wadaj Reservoir 0 15 Inflow Warasgaon Reservoir 0 16 Inflow Mulshi Reservoir 0 17 Inflow Temghar Reservoir 0 18 Inflow Nira Deoghar Reservoir 0 19 Inflow Pawana Reservoir 0 20 Inflow Pimplegaon Joga Reservoir 0 21 Inflow Shirota Reservoir 0 22 Inflow Sina 0 23 Inflow Wadiwale Reservoir 0 24 Inflow Bhogawati-Sina 0 25 Inflow Kalmodi Reservoir 0 26 Inflow Bhatghar Reservoir 0 27 Inflow Dhom Balkawadi Reservoir 0 28 Inflow Mahu DS 0 29 Inflow Koyna Reservoir 0 30 Inflow Kanher Reservoir 0 31 Inflow Urmodi Reservoir 0 32 Inflow Tarali Reservoir 0 33 Inflow Uttarmand DS 0 34 Inflow Morna 0 35 Inflow Wang 0 36 Inflow Warana Reservoir 0 37 Inflow Yerala 0 38 Inflow Kadvi Reservoir 0 39 Inflow Kasari Reservoir 0 40 Inflow Kumbhi Reservoir 0

72 Interim Report Krishna & Bhima River Basins RTSF&ROS

Sr. No. Boundary Type Name Chainage (km) 41 Inflow Dhamani 0 42 Inflow Tulshi Reservoir 0 43 Inflow Radhanagari Reservoir 0 44 Inflow Dudhganga Reservoir 0 45 Inflow Vedganga 0 46 Inflow Chikutra 0 47 Inflow Hiranyakeshi 0 48 Inflow Tamraparni 0 49 Inflow Ghataprabha 0 50 Q-h Krishna 293.03 51 Q-h Vedganga 14.54 52 Q-h Hiranyakeshi 55.225 53 Q-h Ghataprabha 2.71 54 Q-h Tamraparni 64.31 55 Inflow Vasana 0 56 Q-h Dudhganga 69.96 57 Inflow Kera 0 58 Q-h Agrani 42.89 59 Inflow Salpawadi Reservoir 0

8.2.3 Model Outputs The basic outputs of the MIKE 11 hydrodynamic model are discharges and water levels in the main rivers and reservoirs. The model provides additional outputs like flooded area at each cross section, and the total flooded area downstream. Outputs can be obtained for any time steps. Figures 8-7 and 8-8 show sample discharge and water level outputs at daily/hourly and 15 minutes intervals. However, the frequency of the output has to be compatible to the frequency of input data. Hydrographs at daily, hourly and 15 minutes may be produced once the RTDAS provides data at every 15 minutes.

Interim Report 73 RTSF&ROS Krishna & Bhima River Basins

Figure 8-11 Example of Model outputs (Daily water level and discharge at Krishna Ch. 112 km)

Figure 8-12 Example of Water Level Output at 15 minutes interval

8.3 Calibration The calibration of the model has focussed on the routing of river inflows towards the reservoirs. With a very limited number of river cross sections, the calibration of the present pilot model is for demonstration purpose only. Further calibration will be carried out after the surveyed river cross section data become available.

74 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figures 8-13 and 8-14 show sample calibrations against measured discharge and water level at . When the model is used in real time flood forecasting, then the data assimilation compensates for both amplitude and phase errors in routing flood waves. In addition, more detailed data will enable fine tuning the calibration and routing the runoff. This issue will be further addressed in the RTSF&ROS testing and operation phase, when the RTDAS is producing useful and reliable results.

Figure 8-13 Discharge Calibration at Daund

Figure 8-14 Water Level Calibration at Daund

Interim Report 75 RTSF&ROS Krishna & Bhima River Basins 8.4 Reservoir Operation The MIKE 11 Structure Operation (SO) module is being set up to describe the present and future reservoir operation rules for the reservoirs within the Krishna and Bhima river basins. The SO module is applied whenever the flows through spillways and sluice gates are regulated by operation of movable gates or controlled directly as in turbines and pumps. 1. MIKE 11 SO has four control types: 2. Underflow gate (sluice gate) 3. Overflow gate (weirs or inflatable dams) 4. Radial gates (dam spillways) 5. Discharge control (pumps and turbines). The operation rules are applied via a number of logical statements combined with Control Strategies. A Control Strategy is a sequence of commands or rules that determine the way a structure in MIKE 11 shall be operated. A Control Strategy describes how a gate level or a release is set when a Logical Statement Condition is met at a control point. Figure 8-15 shows the Logical Decision.

Figure 8-15 Logical decision tree used in MIKE 11 Structure Operation For a specific structure it is possible to have any number of control strategies by using a sequence of `IF' statements which are evaluated as TRUE or FALSE, consecutively. Each ‘IF’ statement can have any number of conditions that all must be evaluated to TRUE for the `IF'-condition to be evaluated as TRUE. The operation to be applied to a structure is then a relationship between a reservoir water level, or an inflow to a reservoir, or a “day of the year”, or the position of another structure, etc., and the reservoir release conditions defined as gate levels or directly as releases in m3/s. 8.4.1 Implementation of operation rules Even though all the operation rules are being schematised individually for each reservoir, some of the logical statements and corresponding operation strategies

76 Interim Report Krishna & Bhima River Basins RTSF&ROS

are common. To be able to simulate recorded gate levels and releases the first priority is always to look after and apply a time series of gate positions or actual releases. Only then it is possible to simulate accurate reservoir conditions up to the Time of Forecast and to apply releases alternative to the operation rules during the forecast period. The definition of these alternative releases is further detailed in Chapter 9 - Forecasting and Operation. If the SO module fails to locate such time series it steps into the statements describing the official operation rules. These rules are closely linked to the Flood Control Levels (FCL) which again reflect the flood season and the dam safety conditions. The reservoir operation rules have been developed based on daily observations of levels and releases. This is acceptable considering that only a pilot model has been developed but it must be stressed that a final reservoir inflow and flood forecasting model must be based on hourly observations of reservoir levels and releases, surveyed topographical data and calibrated against hourly levels and discharges. Below in Figure 8-16 and Figure 8-17 simulated and observed (daily average) levels and releases are presented for the Koyna and Panchet reservoirs.

Figure 8-16 Koyna Reservoir operation during July 2006

Interim Report 77 RTSF&ROS Krishna & Bhima River Basins

Figure 8-17 Panchet Reservoir operation during July 2006 As releases are closely related to the Flood Control Levels (FCL), these curves are also shown (in red). In both reservoirs an almost perfect match between model results and observations, have been obtained. But it is also known that daily average releases are not applicable when the project reaches a stage where this model is supposed to be used in real-time reservoir operation and optimisation. Before that the reservoir model must be fine-tuned applying real time reservoir observations From Figure 8-17 it can also be seen that after 9 August the operation strategy is slightly changed. Reservoir levels are significantly above the FCL curve but, compared to late July, only a relatively small amount of water is released. This feature is still to be incorporated into the reservoir operation model.

78 Interim Report Krishna & Bhima River Basins RTSF&ROS

8.5 Further Work The following activities are to be carried out in hydrodynamic modelling.  Complete details of all reservoirs and related structures: This task can only be initiated after hourly (real time) reservoir observations become available.  Incorporation of updated catchment runoff after final calibration of the NAM model: This task can only be initiated after hourly (real time) rainfall and reservoir observations become available.  Model Calibration: After all river cross section data are available (from the proposed survey programme) and a DEM has been developed based on these data.  Fine tuning with high frequency real time data: After Data from RTDAS area available  Flood plain modelling and inundation mapping: After all river cross sections and flood plain levels are available from the proposed survey programme  Optimization of reservoir operation: Rule curve optimization during short term operation considering downstream flooding and the need to keep reservoirs as full as possible at the end of the rainy season.

Interim Report 79 RTSF&ROS Krishna & Bhima River Basins

9 FORECASTING MODEL 9.1 Introduction The hydrological model maintains a quantitative memory of the water accumulated in the catchments in the form of soil moisture, and ground water. This accumulated water volume will be released as runoff to the main rivers during the succeeding months, simulated by the hydrological model. Converting the predicted precipitation to runoff hydrographs, the model provides a quantitative response to the predicted weather forecast. The output from the model is fed into the MIKE 11 river model for short term forecasts, and the MIKE Basin water resources model for long term forecasts. Quantitative precipitation forecasts have large uncertainties for extended lead time times, and the runoff becomes correspondingly uncertain when the lead time exceeds a few days. Seasonal forecasts, typically covering the coming three months to one year, rely on probabilistic forecasts of the precipitation from time of forecast to the end of the forecasting period. The probabilistic forecasts are based on the historical precipitation series. Figure 9-1 illustrates the concepts of a long-term forecasting system.

Figure 9-1 Conceptual diagram of long-term forecasting The short term forecasting model is similar to the hydrodynamic model. The forecasting model uses rule operation instead of scheduled releases at the reservoirs, and the data assimilation mode is activated. The setup of the short term forecasting model implies that the model handles both historical data and estimated future inflows and scheduled releases. The period during which historical data are applied is termed the hindcast period, and the period representing the future is termed the forecast period. Figure 9-2 illustrates the concepts and steps of a short-term forecasting system.

80 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 9-2 Illustrations of a short-term forecasting system

Another aspect of forecasting is Data Assimilation, which enables the model to assimilate measured water levels and discharges into the model results during hindcast. Corrections made in order to match simulated and measured results are analysed and used to forecast the error that can be expected in the simulated results for the near future. This error forecast increases the reliability of the forecast results (Figure 9-3).

Figure 9-3 Data Assimilation in RT Forecast

Interim Report 81 RTSF&ROS Krishna & Bhima River Basins

Using data assimilation of discharges in the rivers (or water levels in combination with stage:discharge relations) has two purposes. Firstly, assimilating the data ensures that the correct amount of water is conveyed downstream during the hindcast period. Secondly, the error forecast ensures that the recognised error in the inflow is forecast into the near future, thereby improving the validity of the inflow forecast. In this way the best inflow forecast to the reservoirs are achieved. The required model inputs are:  Operation of reservoirs - rule operated meaning that in order to make forecasts it is necessary to know which rules apply. For the hindcast period measured discharge will be used. During hindcast measured discharges are used. For the forecast period, scheduled releases (or user defined releases) are used. A combined time series will be supplied to the model by the system.  Information for data assimilation and error forecast - comprises measured water levels and discharges that can improve the accuracy of the forecasts. These will be provided in time series (Discharge measurements at model boundaries will be assimilated in the hydrologic model.)  Boundary inflows will be drawn from the database (NAM outputs derived from weather forecasts and the RTDAS) and supplied to the model as time series representing the inflow for both the hindcast and forecast period. The following sections deal with the short-term forecast. 9.2 Overview of the Forecasting and Operation System The real time forecasting and operation system is based on calibrated rainfall- runoff and hydrodynamic models described in Chapters 6 and 8. The system works as a stand-alone Windows application, which does not require in-depth knowledge of complicated models and GIS applications. However, based on a very user friendly interface developed for this project, it is possible to have full control on the on-line forecasting. The system once configured may also run automatically. The forecasting system has also the provision to run different scenarios in offline mode so that comparisons can be made with historical floods forecasted on hindcast mode. The offline mode can also be used during an interactive operational decision making. The setup is an open system in which modifications of key parameters such as forecast locations, time steps, warning levels etc., can easily be incorporated by a trained operator. Figure 9-2 shows the User Interface of the operational forecasting system. The Interface can be used to manage the most common activities in the daily operation of a forecasting system.

82 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 9-4 User Interface for the operational forecasting system The User Interface of the operational system contains (referring to Figure 9-4) the following features:  Left column: Tools available for the operation of the forecasting system:

1. Provision to work in Online or Offline modes 2. Configuration of the operational system (as described in Section 9.3) 3. Import of real time data and Quantitative Rainfall Forecast (QPF) from meteorological forecasts 4. Start of a forecast simulation running as a batch job 5. Export of results to the WEB Portal 6. Modelling Tools to edit and view model setup and results 7. GIS facilities, which can be used to prepare new overview maps 8. Opening the WEB Portal (described in Section 9.7) 9. Scenario Management (described in Sections 9.6)

 An overview map showing the model setup and stations included in the setup. The map includes catchment rainfall (shown as values) and all the real time discharge and Water level stations (shown as coloured squares depending of the station status such as normal, warning and alert levels). The overview map can be changed with more detailed maps as per requirement.

 Graphical and Tabular View of the stations selected when clicking on the overview map (lower right) Interim Report 83 RTSF&ROS Krishna & Bhima River Basins 9.3 System Configuration The Configuration Editor is used for system configuration and for modifications of the forecasting system. The Editor includes all parameters required in the real time operation and presentation of results from the forecasting system. Figure 9-5 shows the Configuration Editor.

Figure 9-5 The System configuration editor Referring to Figure 9-5, following parameters can be specified: 1. Model Folder and Title 2. Forecast time specifications - Time of forecast specified as: YYYY-MM-DD HH:MM or now (real time) - Simulation Period 3. Forecasting locations, location of catchments and reservoirs: Forecast locations and parameters related to the forecasting points are specified in a spread sheet (saved in comma separated text format, which can be opened from the configuration editor). New forecast locations can be included as per requirement when adding new lines to the spread sheet 4. WEB page setup: Specifications of the location for the WEB page, time steps to be included, sizes of plots etc.

84 Interim Report Krishna & Bhima River Basins RTSF&ROS

5. Folder locations: Specification of location for modelling files, batch jobs, maps, GIS and archive. 9.4 On-line operation

9.4.1 Batch Jobs The forecasting activities are controlled from a number of batch jobs combined into an overall batch job, which can run automatically scheduled by the Windows Scheduler. It can also be started start from the Operational Tools as individual activities. Figure 9-6 shows the batch job used for carrying out the forecasting activities with MIKE11 model simulation. It includes four different activities: 1. Preparation of the forecast, 2. Forecast simulation, 3. Reading of results, and 4. Archiving of results.

Figure 9-6 Listing of the batch job for operating a forecast

At the start of the operational forecast job, the operational system includes tools to: 1. Import of data from the real time network (RTDAS) and process the quantitative precipitation forecast (QPF) received from meteorological forecasts, 2. Start the forecast simulation (as shown in Figure 9-6), and 3. Export the results to the WEB Portal. These three operations can also be combined into one overall batch job, which can be scheduled to run automatically from the Windows Scheduler. During high flood periods the overall batch job can run hourly and in other periods it can be scheduled to run on a daily basis. After completion of the combined batch job, the forecasting results can be accessed from the operational platform.

Interim Report 85 RTSF&ROS Krishna & Bhima River Basins

9.4.2 Online tools The Operational system also includes tools which can be used to obtain a more detailed overview of the model simulation. The tools developed include: Mike11 Editor: Overview with access to all model editors and simulation parameters user for the simulation of the forecast as shown in the Figure 9-7.

Figure 9-7 Overview of model editors

MIKE11 log file: Overview of warning and error messages from latest forecast simulation. MIKE11 Result Viewer: Detailed overview of results from the forecast simulation using MIKE VIEW, with a predefined setup specified via the configuration editor. Figure 9-8 shows an example which include the river network, a longitudinal section along a selected river, river cross section at a selected location and time series of discharge and water level.

86 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 9-8 Detailed overview of MIKE11 results from a forecast simulation

9.4.3 Forecasts Using the above described system, forecast of inflows to reservoir and then corresponding outflows have been made on a hindcast mode for the year 2006. Figure 9-9 illustrates the forecast results.

Figure 9-9 Example of inflow forecasts for 6,12,36 and 48 hours

Interim Report 87 RTSF&ROS Krishna & Bhima River Basins

Since the RTSF & ROS is not yet fully operational and there are no river cross sections and flood plain levels in flood prone areas, no water level forecasting has been carried out. While reliable cross sections and water level measurements in the river will await implementation of the RTSF & ROS, the present model uses the water levels in the reservoirs for data assimilation. With at least one season of reliable data, a new data assimilation analysis and testing can be performed, and the most appropriate points for data assimilation identified. 9.5 Reservoir Operation The Reservoir Operation can be performed via the Reservoir Operation Module. The tool is used to show the conditions for each reservoir included in the model setup and it can also be used for scenario simulations (as described in Section 9.6). The page consists of a Grahical View, as shown in Figure 9-10, (water level left axis and inflow & outflow right axis), a Tabular View (timeseries of inflow, outflow and reservoir levels) and a Reservoir Overview on the bottom of the page. Figure 9-10 shows an example for the operation of the Warasgaon Reservoir. The graph shows a 3 days period (1 day before Time of forecast, indicated with a grey vertical line and the forecasting period of 2 days). From Figure 9-10 it appears that the inflow (red line) is higher than outflow (green line) until ‘2006-07-30 03:00’, when the outflow exceeds the inflow. The water level (blue line) increases, therefore in the first part of the simulation, while it decrease when the ouflow eceeds the inflow.

Figure 9-10 Example of Reservoir Operation Module (Warasgaon)

88 Interim Report Krishna & Bhima River Basins RTSF&ROS 9.6 Scenario Management The scenario management tools allow the user to run the forecast model with different data and compare the results from scenario simulations with the original simulation. Simulation of scenarios is activated when running the operation system in an offline mode. After finishing a scenario simulation, the scenario results can be archived, which can be loaded later for further assessment. The scenario management tools also include facilities to disseminate the scenarios to the WEB Portal. As an example, tools for managing two types of scenarios are described below in Sections 9.6.1 and 9.6.2. 1. Rainfall Forecast: which can be used to modify the QPF 2. Reservoir Release: which can be used for operating with user defined releases. In addition to the rainfall forecast and the reservoir operation, the scenario management tools can be used to test changes of other model setup data. As an example, it may be necessary to turn off data assimilation at selected locations and then afterwards test the effect of the modified model setup.

9.6.1 Rainfall forecast scenario The Rainfall Editor can be used to modify the rainfall to a catchment (or a group of catchments) from the original simulation to test alternative rainfall events. The rainfall time series is subdivided into user defined periods, where the rainfall editor provides an overview of the accumulated rainfall within specified periods and has a provision to change the values. In Figure 9-11, periods for the last 24 hours up to time of forecast and the following 12, 24 and 48 hours into the forecasting period can be specified. The 122 catchments in the model setup have been grouped into 18 larger areas with similar rainfall characteristics. The user can now change the original accumulated rainfall with new estimations for each period – each value of the original catchment rainfall time series will then be multiplied with the ratio between estimated and original values.

Figure 9-11 Rainfall forecast editor

Interim Report 89 RTSF&ROS Krishna & Bhima River Basins

9.6.2 Reservoir operation scenarios The reservoir operation module can be used to specify user defined releases from the reservoirs to test alternative reservoir operations and releases of water from the reservoirs. As an example, a user defined release of 500 m3/s has been specified for a period of 12 hours for the Warasgaon reservoir as shown in Figure 9-12. To run this scenario the changes made are saved and a new simulation is started with the user defined release. In Figure 9-12, the figure to the left shows the specification of user defined releases (dark green bar), while the figure to the right shows the result of the scenario simulation (light green curve, calculated spilling).

Figure 9-12 Example of reservoir operation scenario (Warasgaon)

After pressing the refresh button it is possible to inspect the effect of changes on the downstream stations. Figure 9-13 shows a comparison of the water level in the Khadakwasla reservoir (blue curve: original, black curve: with user defined release).

90 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 9-13 Viewing results of reservoir operation scenarios

9.7 Communication WEB Portal All results from the forecast simulations are presented on a WEB Portal. The WEB display of station status takes place via Google Maps will all Google Map facilities like zooming to street level and provision to show data on satellite images, road maps and on terrain maps. From the Google Map it is possible to watch station status at preselected time steps and to select a graphical view of a selected time series clicking on the map. The WEB Page has provision for display of four different data types: discharge, water level, precipitation and data from reservoir (water levels, inflow and outflow)

Two different sizes of WEB pages have been developed:

WEB Page for PC: The PC version (large screen) includes provision for viewing all the four different data types described above. Figure 9-14 shows an example of results presented on Google Map. Each station is coloured according to the actual flood status, for example:

Light blue: Below normal level Dark blue: Above normal level Yellow: Warning level Orange: Alert Level Red: Emergency Levels It is proposed that WRD makes a decision on the different warning and alert levels.

Interim Report 91 RTSF&ROS Krishna & Bhima River Basins

Figure 9-14 The Krishna-Bhima WEB Portal (PC version) for status and forecast Figure 9-15 shows the results presented in a WEB Bulletin for the forecasted discharges along the rivers and inflow forecasts to reservoirs.

92 Interim Report Krishna & Bhima River Basins RTSF&ROS

Figure 9-15 WEB Bulletin showing flow forecasts

WEB Page for Mobile Devices: A compressed WEB Page has been prepared for dissemination of results to mobile devices. The display can be used for Iphone/Ipad/Android or other devices supporting pixel resolutions from 320x480 up to 640x640 pixels. The mobile WEB Page has provision for selecting three different

Interim Report 93 RTSF&ROS Krishna & Bhima River Basins

data types (discharge, reservoirs and rainfall). The upper part of the WEB Page shows the station status presented on Google Maps (including the normal Google Maps features like zooming and selection of different maps). The lower part of the WEB page shows selected graph (when clicking on the map), where it is possible to zoom in and out and to see the data at selected time steps in a tabular view.

Figure 9-.16 shows examples of forecast results in a Mobile WEB Page for the Krishna-Bhima system. The three displays refer to discharge, reservoirs and rainfall.

Figure 9-16 Example from the Mobile WEB Page: left- discharge time series, middle- reservoirs, right- rainfall

94 Interim Report Krishna & Bhima River Basins RTSF&ROS 10 CAPACITY BUILDING

Basic training courses were conducted for the officers of BSD on hydraulics, hydrology and modelling. A suite of modelling software including GIS and user manuals has been installed at BSD with training licenses. However, due to lack of adequate computer facilities and due to the fact that the BSD officers and the consultant have been occupied in collecting and processing data, the training could not be progressed as anticipated. Proposals for study tours and technical training abroad have been prepared and provided to WRD for processing. A high-end database server (DELL PowerEdge T710 High Performance Intel 2S Tower Server with additional hard drive backup facility) has been procured and is being installed at the DHI Project Office. It is proposed to shift the server to the Operational Control Room at Sinchan Bhavan, Pune when the room is ready and after the required infrastructure is developed. Dedicated hands-on-training on the models and the operation system will be conducted after submission of the Interim Report.

Interim Report 95 RTSF&ROS Krishna & Bhima River Basins

11 FURTHER WORK Further work required can be stated in following four groups: (1) data related work by WRD, (2) updating and further development of the modelling systems by the Consultant, (3) implementation of the real time streamflow and flood forecasting and reservoir operation system by the Consultant together with WRD, and (4) Support to capacity development of WRD by the Consultant. Planning of future work is presented in Figure 11-1 (screen of MS Project Manager).

Figure 11-1 Detailed plan of tasks for 2011-2012

96 Interim Report Krishna & Bhima River Basins RTSF&ROS 11.1 Data Related work

11.1.1 River Cross Section Survey The planned river cross section survey by WRD should be completed in time in order to be able to use the models for forecasting during the coming monsoon season.

11.1.2 RTDAS The implementation of RTDAS should be completed well before the start of the monsoon season so that real time data is available to the forecasting models developed. Before implanting the forecasting systems in an operational level, a testing period is required. Therefore, the cross section survey and implementation of RTDAs are two critical paths in the RTSF&ROS project. 11.2 Updating & further development of the models The following activities will be carried out during the next three months:  Complete and improve and the calibration of rainfall-runoff models for all the catchments. Simulation of the remaining seven ungagged catchment will be done by adapting to parameters from nearby similar nearby catchments.  Complete the basin simulation model by further refining reservoir operation rules and using more detailed water use data.  Incorporate the new river cross sections (from survey) in the hydrodynamic model and complete the MIKE 11 model calibration.  Include all reservoirs and refine their description in the model. 11.3 Implementation of the RTSF & ROS Once the real time data becomes available from RTDAS, the forecasting system will be tested and implementation will be initiated in consultation with WRD. A real time reservoir operation guidance will be developed using the streamflow forecast as input. Optimization of rule curve will be developed for selected important reservoirs. 11.4 Support to Capacity Development As described in the Inception Report (Dec 2012), the various capacity development activities will be continued.

11.4.1 On-the-job training The officers of WRD, especially the technical officers of BSD are encouraged to work with the members of the Consultant’s team at their project office. The officers can be involved in all aspects of data processing, model development, test running, calibration, updating and implementation of the forecasting system.

11.4.2 Training courses All the training courses to be given by the Consultant’s team members will be completed as planned. WRD officers are also encouraged to take the recommended training courses at other institutions. The future training courses will Interim Report 97 RTSF&ROS Krishna & Bhima River Basins

be hands-on type so that the officers will be capable of operating the system developed. Officers from other management divisions of WRD are also encouraged to attend the training courses. 11.4.3 Study Tours WRD is requested to complete the approval process of the recommended international study tours and training cum technical visits. 11.4.4 Control Room and Website Once the basic infrastructure of the control room is ready, proposed computer hardware and software will be installed by the Consultant. A high speed dedicated broad band Internet connection should be installed by WRD in the control room. WRD is also requested to register a new domain name of the website in which data and information related to forecasting and reservoir operation will be released.

11.4.5 Program for technical support and strategy for sustainability A programme for the technical support as well a strategy for making the RTSF&ROS sustainable in WRD will be prepared and discussed in future workshops.

98 Interim Report Krishna & Bhima River Basins RTSF&ROS

12 REFERENCES /1/ Contract, RTDSS: HP II/MAHA (SW)/2/2011, INDIA: HYDROLOGY PROJECT PHASE –II, (Loan No: 4749-IN), Consultancy services for implementation of a Real Time Streamflow Forecasting and Reservoir Operation System for the Krishna and Bhima River basins in Maharashtra, 2011. /2/ Technical Offer, Loan No: 4749-IN, RFP No. : HP II/MAHA (SW)/2, Consultancy services for implementation of a Real Time Streamflow Forecasting and Reservoir Operation System for the Krishna and Bhima River basins in Maharashtra, 2011. /3/ Request for Proposal, RFP: HP II/MAHA (SW)/2/, INDIA: HYDROLOGY PROJECT PHASE –II, (Loan No: 4749-IN), Consultancy services for implementation of a Real Time Streamflow Forecasting and Reservoir Operation System for the Krishna and Bhima River basins in Maharashtra, 2011. /4/ DHI (India) Water & Environment, Monthly Progress Reports 1-7, RTSF& ROS. /5/ DHI (India) Water & Environment, Inception Report, RTSF& ROS, December 2011. /6/ DHI Water Environment & Health, User guides MIKE 11, MIKE BASIN, MIKE FLOOD WATCH, 2011. /7/ Water Resources Department, Government of Maharashtra. Documents of various Reservoirs. /8/ National Institute of Hydrology (NIH), Development of Decision Support System for Integrated Water Resources Development and Management, Interim Report, DHI, June 2011. /9/ Government of Maharashtra, Irrigation Department, Dam Safety manual Chapter 7, Flood forecasting, reservoir operation and Gate Operation,1984. /10/ Government of Maharashtra, Irrigation Department, Dam Safety manual Chapter 8, Preparedness for Dealing with emergency situations on dams, 1984. /11/ Government of Maharashtra, Irrigation Department, Dams in Maharashtra, 2000. /12/ Maharashtra Water and Irrigation Commission Report, 1999.

Interim Report 99 RTSF&ROS Krishna & Bhima River Basins

APPENDIX

100 Interim Report Krishna & Bhima River Basins RTSF&ROS

A.1 Sample Satellite Image

Figure A.1 Satellite Image of Krishn-Bhima Basins (Source: RESOURCESAT IRS-P6 AWiFS )

A.2 Status of Reservoir Data

Area - Charact Elevation Guide/Rule Inflo Years of No. Reservoir eristic Frequency Capacity Curve w data Level Table

Bhima Basin Pimpalgaon 1 Y Y Y Y Daily 1970-2009 Joga 2 Manikdoh Y Y Y Y Daily 1984-2009 3 Yedgaon Y Y Y Y Daily 1976-2009 4 Wadaj Y Y Y Y Daily 1985-2009 5 Dimbe Y Y Y Y Daily 1992-2009 6 Chaskaman Y Y Y Y Daily 1991-209 7 Bhama Askheda Y Y Y Y Daily 2000-2009 8 Pawana Y Y Y Y Daily 1976-2008 9 Mulshi Y Y Y Y Daily 1976-2009 10 Temghar Y Y Y Y Daily 2000-2009 11 Warasgaon Y Y Y Y Daily 1987-2009

Interim Report 101 RTSF&ROS Krishna & Bhima River Basins

Area - Charact Guide/Rule Inflo Years of No. Reservoir Elevation eristic Frequency Curve w data Capacity Level Table 12 Panshet Y Y Y Y Daily 1976-2009 13 Khadakwasala Y Y Y Y Daily 1976-2009 14 Ghod Y Y Y Y Daily 1976-2009 15 Ujjani Y Y Y Y Daily 1976-2009 16 Chilewadi Y Y Y Y Daily 2000-2007 17 Andhra Y Y Y Y Daily 2001-2010 18 Kalmodi No Y No Y Daily 2010-2011 19 Wadiwale Y Y Y Y Daily 1989-2008 20 Kasar Sai No Y Y No 21 Sina (Nimgaon) Y Y No Y Daily 2003-2008 22 Sina-Kolegaon No No No No 23 Nazare Y Y No No Daily 2009-2010 24 Gunjawani No Y No Y Daily 2007-2008 25 Bhatghar Y Y Y Y Daily 1976-2008 26 Vir Y Y No Y Daily 1964-2008 27 Nira Deoghar No Y No Y Daily 2000-2008 Nethersole 28 No No No Y Daily 1976-2010 (TATA Andhra) 29 Shirota No No No No 30 Walwan No No No No

Krishna Basin 1 Dhom Y Y No Y Daily 1977-2010 2 Kanher Y Y No Y Daily 1984-2010 3 Urmodi No Y No Y Daily 2001-2006 4 Tarali Y Y Y Y Daily 2002-2009 5 Koyna Y Y Y Y Daily 1961-2012 6 Warna Y Y Y Y Daily 1985-2011 7 Radhanagari Y Y No Y Daily 1976-2005 8 Dudhganga Y Y Y Y Daily 1989-2009 9 Tembhu Barrage No Y No No Satpewadi 10 No Y No No Barrage 11 Dhom Balkawadi No Y No Y Daily 2005-2006 12 Mahu No Y No Y Daily 2004-2006 13 Uttarmand No Y No No 14 Morna(Gureghar) No Y No No 15 Wang No Y No No 16 Kadvi No Y No Y Daily 2001-2009

102 Interim Report Krishna & Bhima River Basins RTSF&ROS

Area - Charact Guide/Rule Inflo Years of No. Reservoir Elevation eristic Frequency Curve w data Capacity Level Table 17 Kasari Y Y No Y Daily 1989-2009 18 Kumbhi No Y Y Y Daily 2001-2010 19 Dhamni No Y No No 20 Tulshi Y Y No Y Daily 1976-2009

A.3 Range analysis of Data (at Proposed RT stations) Discharge Data from G-D Stations

S. Hourly Data Daily Data Stations No Maximum Minimum Average Maximum Minimum Average 1 Amadabad 1869.71 0.00 38.99 1673.12 0.00 15.30 2 Ambeghar 935.09 0.01 25.48 1315.58 0.00 20.85 3 Ankali 3902.92 11.00 281.46 5122.23 0.00 441.07 4 Balinge 1644.21 4.20 133.61 1631.92 31.19 155.69 5 Bubnal 7634.01 0.00 676.95 7604.52 0.00 446.15 6 Dattawadi 3641.88 0.00 33.95 7 Devikavathe 7752.16 0.00 209.33 7602.96 0.00 168.39 8 Ichalkarangi 2906.23 0.01 253.43 2898.24 0.00 196.11 9 Kashti 2603.80 0.00 90.93 2226.35 0.00 22.98 10 Khamgaon 5885.17 1.28 109.76 5200.28 0.00 99.07 11 Late 2389.24 0.00 67.45 1791.53 0.00 26.78 12 Mhaisal 10277.21 0.00 502.96 10195.71 0.00 364.43 13 Nighoje 2834.56 0.00 44.78 2110.92 0.00 27.82 14 Nitwade 1775.13 0.01 110.12 1691.41 0.00 27.82 15 Pandharpur 7771.41 0.00 402.60 16 Pargaon 7939.29 0.00 176.72 8781.81 0.00 125.57 17 Paud 832.54 0.71 57.57 575.81 0.00 12.85 18 PimpaleGurav 1760.56 32.87 1284.65 0.00 28.85 19 Sangali 3062.28 6.02 270.20 3452.29 0.00 329.27 20 Shigaon 2286.23 139.51 2270.02 0.00 99.07 21 Shirur 2743.43 0.00 35.33 1732.23 0.00 22.57 22 Shivade 3434.82 0.00 64.03 2798.12 0.00 70.24 23 SidhewadiSol 2416.14 0.00 9.93 24 Umbre 2017.82 0.00 23.93 1215.54 0.00 18.34 25 Wadange 3041.58 0.00 243.98 2889.64 0.00 156.71

Interim Report 103 RTSF&ROS Krishna & Bhima River Basins

Rainfall Data

Daily Data Hourly Data Daily Stations Maximum Minimum Average Maximum Minimum Average 1 Adkur 280 0 4.79 2 Ajara 233.6 0 5.30 45 0 0.46 3 Akalpa 426 0 10.17 53 0 1.1 4 246.8 0 1.75 5 Ambavade-2 186 0 1.62 60 0 1.01 6 AmbavadeKrd 178.2 0 2.23 7 Ambeghar 260 0 4.22 54.5 0 0.4 8 Amble 207.4 0 3.26 56.7 0 0.29 9 Amboli 286 0 5.24 10 Andhali 110 0 1.21 11 Ardal 200.6 0 3.21 12 Askheda 260 0 2.28 50 0 0.23 13 Atpadi 166 0 1.04 14 Aundhe 390.2 0 4.90 15 Bandalgi 192 0 1.79 58.4 0 0.13 16 Barhanpur 186 0 1.42 65.5 0 0.12 17 Barur 144.6 0 1.71 70 0 0.12 18 Belwade 250.2 0 4.29 60 0 0.32 19 Belwandi 114.5 0 1.18 20 Bhatghar Dam Site 375 0 8.87 21 Borgaon 112.8 0 1.49 33.5 0 0.1 22 Budhawadi 365.6 0 6.38 45.4 0 0.64 23 Chandani 165 0 1.65 24 Chandoh 145.2 0 1.18 25 Chandoli Bk 388.4 0 7.67 26 Chaskman 205.2 0 2.10 58 0 0.15 27 ChichodiP 103 0 1.25 40 0 0.14 28 Chikhalgi 110.3 0 0.93 29 Dajipur 592.4 0 15.69 70 0 1.35 30 Davari 263.8 0 5.13 31 Deoghar 355 0 7.35 60 0 0.55 32 Devikavathe 174.6 0 1.73 100 0 0.16 33 Dhavals 68.59962 0 0.71 34 Dhulgaon 146 0 1.46 60.7 0 0.12 35 Diksal 105 0 1.73 60 0 0.13 36 Gargoti 200.9 0 3.97 37 Gavase 418.8 0 8.86 70 0 0.79 38 Ghisar 376.2 0 9.34 90.4 0 0.77 39 GoregaonW 110.4 0 1.56 40 Gudhe 219.6 0 3.48 91 0 0.28 41 Gureghar 457.2 0 14.10

104 Interim Report Krishna & Bhima River Basins RTSF&ROS

Daily Data Hourly Data Daily Stations Maximum Minimum Average Maximum Minimum Average 42 Harna 0 0 0.00 43 Hatwalan 214 0 1.26 44 Hingangaon 138.6 0 1.53 45 Holkapur 155.4 0 1.96 46 Jambhur 474.2 0 7.63 47 JambreUmgaon 262.2 0 10.50 83.9 0 0.97 48 Jamkhed 175 0 1.88 49 Jangamhatti 341 0 7.52 50 Jawalwadi 140.2 0 1.85 51 200 0 1.31 52 Kabnur 151.2 0 2.71 53 Kadal 115.6 0 2.05 53 0 0.16 54 Kadgaon Har 150 0 1.60 55 Kadus 195 0 1.95 56 Kagal 139.2 0 2.36 57 Kas 402 0 11.58 60 0 0.94 58 KasegaonSan 147.8 0 2.02 59 KasegaonSol 200 0 1.78 60 Kashti 129 0 1.36 84 0 0.12 61 Kasurdi 154.9 0 1.13 62 Kathapur 105.2 0 1.33 63 Katphal 140 0 1.21 64 Tunnel 144 0 2.50 65 Keral 258.8 0 5.14 66 Khamgaon 123.4 0 1.39 60 0 0.11 67 470 0 11.60 60 0 0.94 68 Khed 156 0 1.86 60 0 0.16 69 Kiwale 255 0 2.53 70 K-Mahankal 120 0 1.29 71 Kodapur 490.2 0 7.01 50.4 0 0.59 72 Kolgaon 147 0 1.21 73 Koliye 327.2 0 5.46 74 KoregoanBh 125 0 1.41 75 Kothali 235.4 0 3.27 43.4 0 0.28 76 Kukkadgaon 224 0 1.67 77 Kumbheri 408 0 9.61 55 0 0.9 78 Kurwandi 160 0 2.46 79 Late 144.4 0 1.46 50 0 0.12 80 Lonikand 128 0 1.47 81 Madha 152.4 0 1.65 60.4 0 0.1 82 Mahablesh 550.4 0 17.13 82 0 1.52 83 Mahud 142 0 1.45 84 Malavali 786 0 10.17

Interim Report 105 RTSF&ROS Krishna & Bhima River Basins

Daily Data Hourly Data Daily Stations Maximum Minimum Average Maximum Minimum Average 85 Malewadi 317.4 0 3.19 37 0 0.24 86 95 0 1.02 87 Mandavgan 135 0 1.51 70 0 0.12 88 Mandukali 460 0 11.88 56.5 0 1.06 89 Mangalvedha 200 0 1.59 90 Mangle 350 0 4.69 91 Manikdoundi 146 0 1.42 92 Marali 234.6 0 2.53 93 Mendh 233 0 5.62 94 Mhaisal 103.5 0 1.58 72 0 0.13 95 Miraj 170.4 0 1.71 96 Morgoan 150 0 1.38 97 Morna Colony 230 0 2.70 98 Mulshi 354.7 0 6.40 99 Murum 122.9 0 1.58 100 Nadgadwadi 169.2 0 3.13 68 0 0.26 101 Nagewadi 195 0 3.01 60 0 0.28 102 Nagthane 176.4 0 2.37 103 Nanaj 138 0 1.53 104 Narsinhpur 160.4 0 1.48 105 Nasrapur 321.9 0 3.37 33.7 0 0.28 106 Nighoje 162 0 2.40 56.8 0 0.2 107 Nimaj 115.2 0 1.18 108 Nimgaon G 163.4 0 1.55 50 0 0.13 109 Nitawade 204.7 0 3.34 40 0 0.3 110 Pabal 165.8 0 1.60 111 Padali 393.8 0 10.12 112 Padloshi 250 0 4.73 95 0 0.4 113 Palkhurd 310 0 7.59 45 0 0.41 114 Pangari 351.4 0 10.74 115 Parali 268.4 0 4.40 116 Pargaon 123.19 0 1.35 52 0 0.11 117 Patane 780.6 0 8.95 57 0 0.69 118 Pathari Tank 180 0 1.50 119 Patryachiwad 294.4 0 7.76 72.2 0 0.73 120 Paud 240 0 4.78 60 0 0.46 121 PimpaleGurav 111.8 0 1.89 48.5 0 0.13 122 PimpalgaonJ 215 0 2.94 45 0 0.25 123 Pimpalwandi 142.24 0 1.57 53.8 0 0.13 124 Radhanagari 480 0 12.17 60 0 1.06 125 Rajewadi 167 0 1.38 126 Rakshewadi 105.8 0 1.35 55 0 0.11 127 RanjangaonG 142.4 0 1.56 106 Interim Report Krishna & Bhima River Basins RTSF&ROS

Daily Data Hourly Data Daily Stations Maximum Minimum Average Maximum Minimum Average 128 Rethare 218.4 0 2.58 129 Revechiwadi 590 0 14.35 60 0 1.24 130 Rosa 151 0 1.80 68.5 0 0.14 131 Rukadi 129.6 0 1.78 132 Sagareshwar 96 0 1.69 133 Sakhar 309.8 0 5.09 40 0 0.35 134 Sandavali 397.6 0 11.92 135 Sangola 264 0 1.67 86 0 0.08 136 Sangvi(BK) 191.6 0 3.55 48 0 0.3 137 Sarola Kasar 155 0 1.33 138 Sarud 161.2 0 3.59 60 0 0.32 139 126 0 1.83 140 SatveSavarde 156.2 0 3.04 141 Savale 478 0 8.69 142 Sarvargaon 166.2 0 2.03 143 Shigaon 116 0 2.00 60.7 0 0.16 144 Shikrapur 183 0 1.49 145 Shirala 196 0 2.97 146 Shirgaon 502 0 11.74 70 0 1.3 147 Shirur 117 0 1.39 77 0 0.13 148 Shivade 130 0 2.30 60.5 0 0.17 149 Shive 287 0 2.98 150 ShidhewadiSol 177.4 0 1.74 72.5 0 0.13 151 Sitpur 153.6 0 1.56 152 Songe Bange 181.6 0 2.99 50 0 0.25 153 Supa 180 0 1.25 154 Suratgaon 138.4 0 2.05 155 Surdi 199 0 2.12 156 Surul 241.4 0 6.09 157 Talu Project 137 0 0.99 158 Tandulwadi 2 242.8 0 5.64 159 Tarewadi 157.4 0 3.50 43 0 0.28 160 Targaon 233.4 0 1.78 161 Tekodi 180 0 1.25 162 Tembhurni 170 0 1.57 163 Thapewadi 141 0 2.70 164 Thitewadi 100 0 1.46 165 Thoseghar 363 0 8.53 166 Umadi 113.8 0 1.26 64 0 0.12 167 Umbre 260 0 3.37 56 0 0.3 168 Upshinge 174.8 0 2.02 169 Vinzar 245.2 0 4.43 170 Wadange 210 0 2.74 61.9 0 0.24

Interim Report 107 RTSF&ROS Krishna & Bhima River Basins

Daily Data Hourly Data Daily Stations Maximum Minimum Average Maximum Minimum Average 171 Wagoli 138 0 1.43 46 0 0.13 172 Walwad 127 0 1.25 173 Wathar 175.4 0 1.68 174 Wathar Stn 104.6 0 1.60 175 Wegre 207 0 4.71 34.5 0 0.42 176 Whiram 384.3 0 6.03 177 Yelgaon 210.3 0 3.59 178 Yermala 210.3 0 1.71

108 Interim Report Krishna & Bhima River Basins RTSF&ROS A.4 Updated Meta Data (in Separate Volume)

Interim Report 109