Short answer: Scopus have their own database of publications calcuating citations per year, which is different to the trademarked "impact factor". There are also variations between years and between citescore and 2-year values.
Note on predatory publishers: It is common for predatory publishers to fabricate an impact factor, or come up with some variant of an official impact factor (e.g., "google impact factor"), when they do not have an official impact factor.
However, your case is different. The journal does have an official impact factor (or scopus citescore) which suggests that it probably passes some basic quality threshold. The website is simply listing a different impact factor than what you have found.
While it could be a typo, or even fraud, the most likely answer is that there are many different "impact factors".
Official impact factor: First, I believe impact factor is a trademark traditionally of ISI. But the main point is that "impact factor" in this sense refers to 2-year impact factor using a particular bibliometric database (the Web of Science, I think).
Even here, the impact factor is updated each year and there is a 2-year and a 5-year impact factor that is commonly reported (although without clarification, it should be the 2-year impact factor). So, occasionally, a journal website might be showing a slightly out of date impact factor.
Impact factor more broadly: There is a boarder concept of impact factor that relates to the general formula of dividing citations by publications over a given period. There are subtle ways that this formula can be tweaked based on what counts as a publication. In particular, the choice of database dramatically influences what publications contribute towards citation counts. The more inclusive the database, the more citations a given publication will receive, and the higher the "impact factor" will be.
The scopus metric is typically called citescore. I think the scopus metrics focus on a three year period and uses a different database (i.e.,Scopus rather than web of science) for generating citation counts. https://journalmetrics.scopus.com/ . Also see http://www.scimagojr.com/ for another set of rankings also based on Scopus.
In general, the original ISI impact factor has a certain influence in the market. Various reasons include quality control of their database of citations and the fact that they were the first big player in this space. They have also trademarked the term "impact factor", which is interesting given that the concept of dividing citations by publications is a pretty straightforward idea.
Nonetheless, the meaning of an index of citations per year is often based on benchmarks within a field. For example, researchers in a field get to know the impact factors of the journals in a given field and they form rules of thumb about what is poor, good, and great. In this sense, if you change the index, then the rules of thumb may change. For example, if you used Google Scholar to calculate impact factors, then impact factors would probably increase massively (e.g., by a factor of 2, 3, or more more), because Google Scholar is highly inclusive. So for example, a bunch of journals that had an ISI impact factor or 2, now has a GS impact factor of 6 or 7. So some care is required when using a new index to ensure that relative benchmarks are recalibrated.
Check out this review of the mapping of scopus citescores with ISI impact factors.