Notes:

1) The position of the buoy was not correctly put into the code.  This has been rectified
'TJS Update: Correct position of L4 buoy
'TJS This to Determine when buoy out of position
Const lat_base = 50.25
Const lon_base = -4.217

2) The Met Station (which is also used for the GPS position) requires the string which is output from the GMX to be consistent:

Line 164 - 195
Function wind_measurement 
    SerialInRecord (COMU7,trame_vent,&H2,0,&H3,nbr_vent,00)
    SplitStr (trame_vent_split(),trame_vent,",",11,4)
    wind_speed_digital = trame_vent_split(2)
    wind_direction_digital = trame_vent_split(3)
    pressure_digital = trame_vent_split(5)'OU 5
    humidity_digital = trame_vent_split(6)'OU 6
    temperature_digital = trame_vent_split(7)'OU 7
    dew_point = trame_vent_split(8)'OU 8
    GPS_pos=trame_vent_split(9)'OU 9
    SplitStr (GPS_split(),GPS_pos,":",2,5)
    Latitude_decimal=FormatFloat (GPS_split(1),"%9.6f")
    Longitude_decimal=FormatFloat (GPS_split(2),"%9.6f")
    WS_voltage=trame_vent_split(11)'OU 14
  
      Units wind_speed_digital = m/s
      Units wind_direction_digital = °
      Units pressure_digital = hPa
      Units humidity_digital = %
      Units temperature_digital = °c
      Units WS_voltage = V
  
      lat2 = Latitude_decimal * PI / 180
      lon2 =  Longitude_decimal * PI / 180
      lat = lat_base * PI / 180
      lon = lon_base * PI / 180
      $dlo = (lon2 - lon) / 2
      $dla = (lat2 - lat) / 2
      tempo1 = (SIN($dla) * SIN($dla)) + COS(lat2) * COS(lat) * (SIN($dlo) * SIN($dlo))
      distance = ((2 * ATN2(SQR(tempo1), SQR(1 - tempo1)))*6378)*1000

End Function

Without sight of the GMX output string it is difficult to verify if either the string is being parsed correctly or there is a problem 
with the GMX500

3) The GMX500 is connected to COMU7

Line 1334
  SerialOpen (COMU7,19200,0,0,100) 'GMX500 + GPS

However there don't appear to be commands to either close or flush the serial ports - which may be good practice. 

4) Lines 1299 - 1318

Possible issues with data averaging ingesting corrupted data over the whole 10 minute period
Have therefore changed the data interval for Met and GPS to be the preceding 5 minutes

DataTable (Meteo_avgd,True,-1 )
' TJS change the data interval to be for preceding 5 minutes
'  DataInterval (0,10,Min,10)
  DataInterval (5,10,Min,10)
  Average (1,temperature_digital,FP2,False)
  Average (1,pressure_digital,FP2,False)
  Average (1,humidity_digital,FP2,False)
  WindVector (1,wind_speed_digital,wind_direction_digital,FP2,False,0,0,1)
  FieldNames ("wind_speed_digital, wind_direction_digital")
  Average (1,WS_voltage,FP2,False)
EndTable

DataTable (Position,True,-1 )
' TJS change the data interval to be for preceding 5 minutes
'  DataInterval (0,10,Min,10)
  DataInterval (5,10,Min,10)
  Sample (1,Latitude_decimal,FP2)
  Sample (1,Longitude_decimal,FP2)
  Sample (1,distance,FP2)
EndTable

5) Another option could be to remove data averaging altogether and use the following:

DataTable (Meteo_avgd,True,-1 )
  DataInterval (5,10,Min,10)
  Sample (1,temperature_digital,FP2)
  Sample (1,pressure_digital,FP2)
  Sample (1,humidity_digital,FP2)
  Sample (1,wind_speed_digital,FP2)
  Sample (1,wind_direction_digital,FP2)
EndTable