From the observed frequency dependence of surface n»sn-!r.a<le radio rioi>*<a 

 Skot&sl/ haa shown that the tturface noiae povnr function, P fi C t , d ) , e/prtsaed 

 in units of power, watts par bandwidth, displays an inverse dt!p«tr!<ti»nc»' upon 

 f requpr,cy, f, and distance, 6. When expressed xn decib*la relative to 1 raW 

 p«r detection bandwidth, b, (daro/b) 



P s (f.d) - E, + t 2 <d-k) + E 3 (6-k) 2 , (1) 



where It It d constant which when oet to 2.5 ctiloa providea a good representa- 

 tion of vhf and uhff surface radio noiea data. Coefficients E. {%. « a , + bif, 

 i • 1/ 2 and 3) In ths above equation contain the frequency c«!p<$.id«nca of the 

 surface noise power function aid are determined frora three conetramtn placed 

 on the above equation at a specified frequency, f. 



1. At the point d «* C the derivative of P with respect to d, 



SUo ■ •< 



a condition that haa been observed to occur for composite urban man-made radio 

 noise data. 



2. At d - 2.5 miles the noise power function at frequency f in MHz 

 equals a least-squares-regression iine derived froa; business zone data in 

 dBm/b, 



P a - -89.9 - 12.3 log f B , «jd " (3) 



3. At d = 10 miles the noise power function at frequency f equals a 

 least-sqir res regression line derived f rem residential zone data in dBm/b, 



P w " -98. S - 12.7 loa f_. (4) 



Using these constraints to 60lve equation (1) for coefficients E. , g_ an d E 3 

 gives 



E, - P a# IS) 



K 2 - 5 E 3 # and (6) 



"3 " < p b " Pa^ 93 - 75 ' (7) 



