The following are the TC SIGMETs issued by HKO last year (no TC yet this year, we are also cold here) which cover FCST (forecast to enter FIR) and OBS (observed to be in FIR) scenarios.  They basically differ in the issue and valid times (FCST -> different but OBS -> same):

WCSS20 VHHH 100515
VHHK SIGMET 3 VALID 100630/101230 VHHH-
VHHK HONG KONG FIR TC SARIKA FCST AT 0630Z N2030 E11706
CB TOP FL400 WI 120NM OF CENTRE MOV NNW 11KT NC
FCST 1230Z TC CENTRE N2142 E11642=

WCSS20 VHHH 100810
VHHK SIGMET 5 VALID 100810/101400 VHHH-
VHHK HONG KONG FIR TC SARIKA OBS AT 0800Z N2024 E11700
CB TOP FL400 WI 120NM OF CENTRE MOV NNW 11KT NC
FCST 1400Z TC CENTRE N2124 E11636=

The above examples remind me the difficulties we have when there are more than one TCs in or forecast to be in our FIR.  The current format only allows the description of one TC in a single message so we have to issue two at the same time if required.  However, the WMO GTS header + datetime group enforces uniqueness of messages (i.e. WCSS20 VHHH 100810 should be unique).  Some centers solve this problem by reserving a number of headers for multiple TCs, for example WCSS20, WCSS21, WCSS22, etc but since HK has only one defined header we do this by issuing bulletins with a time difference of one minute, i.e. WCSS20 VHHH 100810 and WCSS20 VHHH 100811, to honour the uniqueness requirement.

I did not see the WMO header being modeled in our LDM.  Perhaps that has already been dealt with somewhere else?