For the word 九日 (ここのか{LHHH}), both OJAD and weblio agree on the pitch accent of the word itself, but for the particle that comes afterwords (say 九日の), they differ by this:
OJAD: ここのかの{LHHHH}
weblio: ここのかの{LHHHL}
From what I can see both sources seem fairly reliable, so does this mean that both pitches are considered standard?
I personally read 九日から like ここのかから【LHHHHH】, but ここのかから【LHHHLL】 is also acceptable. 九日の天気 is read both like ここのかのてんき【LHHHHHLL】 and ここのかのてんき【LHHHLHLL】, but I feel the former is more common.
Some words are pronunced in two ways depending on the speaker. See: ズボン pronunciation variation & explanation for and Are there any rules to the intonations they are discussing in this video?
FWIW, Google Translate agrees with me.
Sometimes a word does have multiple "valid" pitch accent patterns. Daijirin often lists multiple pitch downstep numbers for terms that have them, like, say, the entry for とらまえる, which lists patterns 4 and 3. That said, the Daijirin entry for ここのか only lists pitch pattern 4, with a downstep after the fourth mora.
Checking in my local copy of the NHK アクセント辞典, they list two patterns -- 4 for the noun, and 0 (no downstep) for the adverb. However, a bit confusingly, both entries (noun and adverb) include the same sample sentence:
あとココノカデ{LHHHL}期限が来る (noun)
あとココノカデ{LHHHH}期限が来る (adverb)
It seems that Daijirin wound up using one pattern, and OJAD the other, but if the NHK is anything to go by, it looks like both patterns 4 and 0 are recognized as correct.
