| 1 | = PDAF_correlation_function = |
| 2 | |
| 3 | |
| 4 | This page documents the routine `PDAF_correlation_function` of PDAF. It was introduced with PDAF V2.3. |
| 5 | |
| 6 | The routine provides the value of a chosen correlation function. It can be used, e.g. to generate non-diagonal observation error covariance matrices. |
| 7 | |
| 8 | The interface is the following: |
| 9 | {{{ |
| 10 | SUBROUTINE PDAF_correlation_function(ctype, length, distance, value) |
| 11 | }}} |
| 12 | with the following arguments: |
| 13 | * `ctype` : `integer, intent(in)`[[BR]] Type of weight function:[[BR]] (1) Gaussian function scaled to f(0)=1 [[BR]] (2) 5th order polynomial with support radius sradius (Gaspari&Cohn 1999; 0 for distance>sradius) |
| 14 | * `length` : `real, intent(in)`[[BR]] length scale:[[BR]] for (1) standard deviation[[BR]] for (2) cut-off length (f=0 for distance>=length) |
| 15 | * `distance` : `real, intent(in)`[[BR]] Distance at which the function is evaluated |
| 16 | * `weight` : `real, intent(out)`[[BR]] Function value |
| 17 | |