After GPS, ISRO Is Now Working On World's Most Efficient Weather Prediction Satellite!

[ashok321]

The Indian Space Research Organisation (ISRO) is developing a new weather-forecasting satellite to predict the beginning of cyclones in oceans. About 300 scientists are working on the satellite named ScatSat-1, at the Space Applications Centre (SAC) in Ahmedabad.

The amazing fact about this satellite, weighing 301 kg, is that it is being built at 60% of the actual cost, and in one-third of the estimated time. About 40% of satellite is made by recycling the leftover equipment from previous satellite missions. It will be launched in July 2016 and will have a mission life of five years.

OceanSat-2, which was launched in 2009 and was known for its accurate predictions of cyclones like Hudhud and Phailin, will be replaced by ScatSat-1. Although OceanSat-2 stopped functioning in February 2014, currently, ISRO is getting most of the weather information from INSAT-3D satellite.

“Normally, it takes about three years to build a satellite of this class from scratch. However, as we have sourced 40% of the parts used in ScatSat-1 from spares of previous missions, we will complete it in a year’s time,” said Tapan Misra, director of the SAC.

At the time of launch, ScatSat-1 will be a piggy ride with another satellite, and this will help save costs even further.

A scatterometer in the satellite will help predict formation of cyclones in waterbodies like seas and oceans. Predictions such as these will help in timely evacuation and minimise human casualties. A scatterometer measures the scattering effect produced while scanning the surface of the Earth from an aircraft or a satellite. It measures the direction and speed of winds over the seas and oceans. This satellite has been designed to withstand multiple system failures, unlike OceanSat-2.

ScatSat-1 is a polar orbiting satellite and will take two days to cover the globe. The data will be used by organisations like NASA, European Organisation for the Exploitation of Meteorological Satellites, and National Oceanic and Atmospheric Administration as well.

 

[TimeTraveller]

The Indian Space Research Organisation (ISRO) is developing a new weather-forecasting satellite to predict the beginning of cyclones in oceans. About 300 scientists are working on the satellite named ScatSat-1, at the Space Applications Centre (SAC) in Ahmedabad.

The amazing fact about this satellite, weighing 301 kg, is that it is being built at 60% of the actual cost, and in one-third of the estimated time. About 40% of satellite is made by recycling the leftover equipment from previous satellite missions. It will be launched in July 2016 and will have a mission life of five years.

OceanSat-2, which was launched in 2009 and was known for its accurate predictions of cyclones like Hudhud and Phailin, will be replaced by ScatSat-1. Although OceanSat-2 stopped functioning in February 2014, currently, ISRO is getting most of the weather information from INSAT-3D satellite.

“Normally, it takes about three years to build a satellite of this class from scratch. However, as we have sourced 40% of the parts used in ScatSat-1 from spares of previous missions, we will complete it in a year’s time,” said Tapan Misra, director of the SAC.

At the time of launch, ScatSat-1 will be a piggy ride with another satellite, and this will help save costs even further.

A scatterometer in the satellite will help predict formation of cyclones in waterbodies like seas and oceans. Predictions such as these will help in timely evacuation and minimise human casualties. A scatterometer measures the scattering effect produced while scanning the surface of the Earth from an aircraft or a satellite. It measures the direction and speed of winds over the seas and oceans. This satellite has been designed to withstand multiple system failures, unlike OceanSat-2.

ScatSat-1 is a polar orbiting satellite and will take two days to cover the globe. The data will be used by organisations like NASA, European Organisation for the Exploitation of Meteorological Satellites, and National Oceanic and Atmospheric Administration as well.

Nice.......

 

[SrNair]

@SOHEIL

 

[SOHEIL]

 

[subincb]

I guess the problem is with our simulation models. I think we need to work on that to perfect our forecast. Not an expert. Experts please comment.

 

[thinkingcap81]

I guess the problem is with our simulation models. I think we need to work on that to perfect our forecast. Not an expert. Experts please comment.

Hi,

You are absolutely right. Due to serious lacunae in our understanding of actual phenomena, climate models give grossly inaccurate results. Weather models are better because their predictions are, by design, valid only for a few days.

All over the world almost all climate models perform poorly in understanding the Indian climate system.

Climate models are good to understand certain 'trends', but take the results with a pinch of salt.

Even if we could eventually understand all the phenomena and model them correctly, we cannot give long term 'accurate' predictions because of something called 'sensitive dependence on initial conditions'.

-------------------

I don't like the name ScatSat .

 

[Providence]

To model a weather phenomena, you need heavy usage of partial differential equations whose solutions might not even exist (other than extrapolation). A precise modelling forget practically, even mathematically would be very difficult.

 

[Papa Dragon]

DRDO/HAL should also work on the lines of ISRO and change it's administration accordingly. No wonder most of ISRO's projects are a success.

 

[thinkingcap81]

one more so called expert who does not know fluid mechanics misguiding others.
the answer is there is no way 3 dimensional fluid flow can be predicted accurately, when there is boundary separation. all 3 dimensional fluid flow model give approximate result as there are many variables to calculate the flow. hence weather forecast throughout the world even USA can go wrong.

I don't understand why you are saying this. Given enough computational power and a grid size that resolves till the Kolmogorov scale, why can't Direct Numerical Simulations of the Navier-Stokes equations predict flow separation with high enough precision?

To model a weather phenomena, you need heavy usage of partial differential equations whose solutions might not even exist (other than extrapolation). A precise modelling forget practically, even mathematically would be very difficult.

High precision (may not be accuracy) of the solution of the PDEs may be possible, given enough computational power. However the PDEs themselves may be wrong and not represent the actual physical phenomena.

If the PDEs correctly represent physical phenomena, then given realistic initial and boundary conditions they should have solutions - because the actual physical phenomena does occur.

 

[KingRaj]

When will ISRO launch CCI SAT?

 

[Providence]

High precision (may not be accuracy) of the solution of the PDEs may be possible, given enough computational power. However the PDEs themselves may be wrong and not represent the actual physical phenomena.

If the PDEs correctly represent physical phenomena, then given realistic initial and boundary conditions they should have solutions - because the actual physical phenomena does occur.

So? Wind has been blowing on earth for last 3 billion years. Does wind needs to give a **** about the number system invented by humanoids who came only 1 million years ago ?

 

[egodoc222]

We need supercomputers for data analysis and simulation of weather data!

 

[thinkingcap81]

I thought you will not post further in this forum again. after climate model and weather model cock and bull story. After more than 1 day of research you came back with a reply which is way of the mark. Did you even understand what I had posted?. I was talking about 3d fluid flow and you jumped with fluid separation. If Euler had read your post he would have come after you. this is what happens when you do not know the subject and you comment on it. It is okay to make mistakes and one should be gracious enough to accept it.

What is 'boundary separation'? Are you talking of 'boundary layer separation'?

When you spoke of 'models' did you mean RANS and/or LES codes? These codes will not predict all flow situations because they have 'rational approximations' built into them for predicting many (not all) flow situations.

Why can't separation in 3D fluid flow be modeled with high fidelity if we have good DNS codes with the required high grid resolution? Please explain.

My point is very simple - our understanding of the climate system is very incomplete and so our equations do not take into account all of them - and hence the difficulties in predicting future climate. Separation is a very minor part in lack of predictability of our climate and weather models (weather forecasts are usually reliable for a few days). Show me a paper in a good peer-reviewed journal which endorses your contention that climate and weather prediction is hampered primarily because of 'separation'.

Even if we land up with the correct set of equations and the required amount of computational power, at best we will get long-term predictions that are accurate in the statistical sense - achieving this itself is a herculean task.

Or am I missing what you are saying?

So? Wind has been blowing on earth for last 3 billion years. Does wind needs to give a **** about the number system invented by humanoids who came only 1 million years ago ?

Correct! I cannot disagree with this point!

I said that if we have a correct set of equations that claim to 'model' the weather and climate system correctly, then given correct initial and boundary conditions they would have a solution - because by their claim they need to give answers that resemble real-life scenarios.

Of course what you said 'A precise modelling forget practically, even mathematically would be very difficult' is true because of the present understanding and computational power at our disposal.

 

[thinkingcap81]

It seems you got some help to understand about 3d fluid flow.
lets recap what you learnt from my post
1.3d fluid flow cannot be predicted accurately because all 3 dimensional fluid flow model give approximate result as there are many variables to calculate the flow.
2.mathematical model is available for steady flow with many initial condition like no boundary layer separation, viscosity,density etc constant. I mentioned boundary layer separation as for that condition there are no precise mathematical model with viscosity changing between layers.
do not limit yourself with only boundary layer separation. there are so many factors like unsteady flow, different fluids like water,air,etc interacting. for your information.
do you know that for predicting you need sensors some are placed in various parts of ocean that measure precise temp,pressure, density etc. sensors do not cover entire ocean. Even with sensors with various fluids interacting, super computer with various gpus
give approximate result and prediction given for 24 hours have gone wrong. i can go on writing giving many reason but one fundamental reason is that it is a 3d flow of fluid. Hope you will remember this lesson and the next time some one asks you why weather prediction is difficult, you can say it is because of 3d fluid flow with various variables is difficult to predict mathematical.

There are many things in fluid dynamics that i don't know. But you have not given me any information that I don't know.

Replying to post #5 i said that we are not in a position to model atmospheric and oceanic phenomena because we do not understand many physical processes and so many of the equations contain only part of the information needed. Then i said that even if we have a complete set of correct equations we need a computational facilities which are not available presently, and even if we did we would be hampered by lack of accurate observational data to feed into the model as initial (and boundary) conditions.

To this you replied in post #7 that we cannot model weather and climate accurately because 3D flow cannot be predicted primarily because of something you call as 'boundary separation'.

Since then i have been asking you for proof why 3D flow cannot be modeled with high fidelity if we have proper knowledge of the flow parameters, high grid resolution and 'super-super computers'. You have not answered this point.

Lets make it very simple - take 3D incompressible fluid flow, fix the fluid parameters, use only continuity and momentum equations, and please tell me why we cannot predict flow evolution which is very reliable in the statistical sense. Steady flow, unsteady flow whatever. Even 3D transient flow can be modeled well with DNS. To do this we of course need high grid resolution and computational power that is not available today.

You are talking of mathematical models, i am talking of DNS at high resolution. Models are employed if they have rational approximations that mimic actual fluid flow such that their predictions are reliable. We use such models because it saves significantly on computational effort. No wonder such models are unreliable over a variety of flows for which they have no rational approximations.

Hence DNS at high resolution which uses computational brute force to solve the equations governing fluid flow.

In post #16 I also asked you to give me solid proof which endorses your contention that climate and weather prediction is hampered primarily because of 'separation'. You have not given me that.

Now you come and say what i have been saying - in your words 'do you know that for predicting you need sensors some are placed in various parts of ocean that measure precise temp,pressure, density etc. sensors do not cover entire ocean'. In short you endorse my point that we do not have accurate information and not your original point about 'boundary separation'. And show, as you said, that this lack of information leads to incorrect prediction of 'boundary separation' and that this 'boundary separation' is the most important cause for weather and climate unpredictability.

3D fluid flow is more difficult to model than 2D flow, which sane person will disagree?

 

[Not Sure]

@thinkingcap81 , Thank you for your well organized and highly informative posts, they sparked my interest in fluid mechanics. I would like to learn a lot more about the subject, and if you could suggest me some books on the subject that could help an ignorant like me to understand how to predict fluid behavior, I'd be grateful to you.

P.S. Why even bother replying to condescending time wasters like that @asiaticlion in here?