normality of multi-regression analysis error (sig. = 0 and bigger than =0.05)

#1
Dear member of talkstats:
I have tried to estimate solar radiation value (dependent variable ) and using multi linear regression analysis. My independent value is relative humidity , cloudless , soil temperature.
I have 2 type of data which are monthy (data size around 140) and daily (data size 730).
I have no normality problem in mouthly estimation but I have problem with daily.

What I need to do now ? Can I use daily regression analysis and normality test in the article ?
What I need to report ? is it correct to use in the estimation solar radiation with the daily multi regression model ?

for daily normality of sig. is zero

monthly is bigger that 0.05
i also add picture for daily.
 

Attachments