Another good ol' segmentation error in fortran - fortran

Question

I know that people ask a LOT of segmentation err questions here, but I've put my effort in solving this problem for more than three hour and still wasn't able to solve this. :/ So here is my code:
c sinle event analysis
implicit real(a-h,o-z)
real day(12), nmonth(12), year(12), clas(12),
$ hour(12), nmin(12)
integer mark(12)
real tst(12), D(12), avgP(12,6), avgA(12,6)
integer k, m, n, g
real time(2054904), proa(2054904), prob(2054904), w1(2054904),
$ w2(2054904), w3(2054904), w4(2054904)
D(1) = 31; D(2) = 28; D(3) = 31; D(4) = 30; D(5) = 31;
D(6) = 30; D(7) = 31; D(8) = 31; D(9) = 30; D(10) = 31;
D(11) = 30; D(12) = 31
open(100,file='singleE.dat')
do i=1, 12
tst(i)=0
enddo
900 do i=1, 12
read(100, 1150) day(i), nmonth(i), year(i),
$ hour(i), nmin(i), clas(i)
do j=1, 12
if (int(nmonth(i)).EQ.(13-j)) then
tst(i) = tst(i) + D(12-j)
nmonth(i) = nmonth(i)-1
endif
enddo
tst(i) = tst(i) + day(i) + (year(i) - 2010)*365
$ + (hour(i) + nmin(i)/60)/24
if (year(i) > real(2011)) then
tst(i) = tst(i) + 1
endif
enddo
open(200,file='hole.dat',status='OLD')
k = 0
do i=1, 2054904
read(200,950) time(i), proa(i), prob(i),
$ w1(i), w2(i), w3(i), w4(i)
enddo
mark = 0
do i=1, 12
do j=1, 2054904
k = k + 1
if(abs(tst(i)-time(j))<0.0001) then
mark(i) = k
endif
enddo
enddo
n = 5;
do i= 1, 12
do j= 1,6
avgP(i,j) = 0
avgA(i,j) = 0
enddo
enddo
do i=1, 12
if (mark(i).EQ.0) then
go to 750
endif
do j = (mark(i)-(n+1)*1440), (mark(i)-n*1440)
avgP(i,1) = avgP(i,1) + proa(j)
avgP(i,2) = avgP(i,2) + prob(j)
avgP(i,3) = avgP(i,3) + w1(j)
avgP(i,4) = avgP(i,4) + w2(j)
avgP(i,5) = avgP(i,5) + w3(j)
avgP(i,6) = avgP(i,6) + w4(j)
enddo
do g = (mark(i)+n*1440), (mark(i)+(n+1)*1440)
avgA(i,1) = avgA(i,1) + proa(g)
avgA(i,2) = avgA(i,2) + prob(g)
avgA(i,3) = avgA(i,3) + w1(g)
avgA(i,4) = avgA(i,4) + w2(g)
avgA(i,5) = avgA(i,5) + w3(g)
avgA(i,6) = avgA(i,6) + w4(g)
enddo
750 print *, avgP(i,1), avgP(i,2), avgP(i,3), avgP(i,4),
$ avgP(i,5), avgP(i,6)
enddo
850 close(i)
950 FORMAT(F12.7,2x,E10.3,2x,E10.3,2x,E10.3,2x,E10.3,
$ 2x,E10.3,2x,E10.3)
1150 FORMAT(F2.0,1x,F2.0,1x,F4.0,1x,F2.0,1x,F2.0,4x F3.1)
end
The part that is causing me trouble is the loop here:
do i=1, 12
if (mark(i).EQ.0) then
go to 750
endif
do j = (mark(i)-(n+1)*1440), (mark(i)-n*1440)
avgP(i,1) = avgP(i,1) + proa(j)
avgP(i,2) = avgP(i,2) + prob(j)
avgP(i,3) = avgP(i,3) + w1(j)
avgP(i,4) = avgP(i,4) + w2(j)
avgP(i,5) = avgP(i,5) + w3(j)
avgP(i,6) = avgP(i,6) + w4(j)
enddo
do g = (mark(i)+n*1440), (mark(i)+(n+1)*1440)
avgA(i,1) = avgA(i,1) + proa(g)
avgA(i,2) = avgA(i,2) + prob(g)
avgA(i,3) = avgA(i,3) + w1(g)
avgA(i,4) = avgA(i,4) + w2(g)
avgA(i,5) = avgA(i,5) + w3(g)
avgA(i,6) = avgA(i,6) + w4(g)
enddo
enddo
Using gdb, I found out that the 'j' loop is causing trouble. All of the parameters are fine, but every time I execute the program the 'j' loop goes only once. The peculiar thing is that as 'i' increases the processes start to cripple one by one. for example, at i = 1 the loop executes well. Then, at i = 2, avgP(i,6) = avgP(i,6) + w4(j) causes the seg fault. At i = 3, avgP(i,5) = avgP(i,5) + w3(j) causes the seg fault and in the end at i = 7, entire loop doesn't work. What a strange error! Hope that I can get some help with this.

Answer 1

The loop
do j=1, 12
if (int(nmonth(i)).EQ.(13-j)) then
tst(i) = tst(i) + D(12-j)
nmonth(i) = nmonth(i)-1
endif
enddo
potentially tries to access the value D(0) when j=12 however D is dimensioned 1:12 so who knows what is being written to tst and its consequences.
This should be caught by turning on bounds checking.