Yes, it is necessary, LLC, IPS, Fisher tests can help to identify the unit root in your data.

stationarity test are necessary for panel, if you make mutliple regression.

But, you could also make a cointegration relationship if the variables are integrated with first order. I(1)

You should do the unit root test first. If the null  hypohtesis of unit root  isn't rejected, co-intergration analysis is needed; otherwise, regression is enough.

ok . thanks

kindly suggest do we apply stationarity test for panel data if we require SURE MODEL