Let's start with a table of two possible uniscalar, decimal, metric calandar systems, one whose nomenclature is based on an inventive portmanteau of “d” for decimal and the word “time”(http://serg.us/dime/); the other on a more standard and familiar combination of Greek/Latin prefixes and a Greek root meaning “time” (think: chronological, chronometer, etc.):
“dime” names 
“chron” names 
definition 

dillenium 
gigachron 
100000 days = millenium 
decade 
megachron 
10000 days = decade 
dear 
kilochron 
1000 days = year 
donth 
hecachron 
100 days = month 
deek 
decachron 
10 days = week 
day (or dime) 
chron 
1 day 
dour 
decichron 
1/10 of a day = hour 
dinute 
centichron 
1/100 of a day = minute 
decond 
millichron 
1/1000 of a day = second 
In the above chart, some license is taken with the Greek prefixes giga and mega, which normal refer to multiples of a billion and million, respectively, but in this case, refer to incremental powers of 10.
In this decimalbased metric calendar system, the day / chron would serve as the pivotal unit and could relate analogically and intuitively to the time it takes the earth to rotate on its axis, which is naturally changeable and would require periodic adjustment of all the absolute values on the scale.
Alternatively, the decond / millichron could be the pivotal unit and relate digitially to a constant, atomic measurement, in which case the units of time would remain fixed but the “day” would be mostly out of sync with the daylight/darkness and tidal cycles due to the variable lengths of time it takes the earth to rotate.
In either of these cases, the dear / kilochron ( = year ) would be totally out of sync with the revolution of the earth around the sun where seasons are defined as fractions (usually quarters) of that revolution. Lunar cycles also would not line up in any way with the donth / hecachron (not that they do it that well with months in our nondecimal Gregorian calendarand who really cares?).
It is doubtful any modern society would now accept a week/ deek/ decachron as being 10 days, since the week is part and parcel of the modern world's current notion of how long we work before having one or two days off for rest. However, if the new notion included the concept of a 5day weekend juxtaposed to a 5day work week, then the level of acceptance might go up, at least by yours truly!
It is also doubtful that any metric calendar system would be palatable if the day did not relate closely to the rotation of the earth on its axis. A solarignorant day would just be too unintuitive and serve little purpose. Any other scalar units could be unrelated to natural events, though not have a unit to “celebrate” the completion of one revolution of the earth around the sun and its concomitant cycling of weather and seasonal patterns would seem rather unpopular. Think about it: “New Dear's Eve” would end up being wintry one time and springy the next or even fallish or summery the next. One could argue that in our current system, Aussies are toasting the New Year in summer while New Yorkers are drinking bubbly in winter. However, on “New Dear's” everyone around the planet would be tooting their horns in a different season every dear. Maybe it's of little consequence for that particular holiday, but what about holidays like Easter that celebrate fertility and the rebirth of crops or Halloween the harvesting of those crops? Yes, such seasonal feasts are scheduled in different parts of the globe according to regional climates and customs, but in most calendars, the dates for these events don't shift too dramatically as they would in a dime system. Cyclical, seasonal feasts typically relate to star, earth , and planting phenomena and not to arbitrary numerical increments; for them to do otherwise, would be meaningless.
Frankly, a decimal metric system for calendar time makes sense only in the minds of digital computers (Unix, for instance) which don't care about things like the rising of the sun, the falling of leaves, or the length of a work week, or how tiny a nanosecond is. Various metric systems are already in place for computational use, as well they should. Humans, though, have never accepted a uniscalar calendar in normal commerce for more than a short political time (even Napolean cancelled his own scheme) and never will on a permanent basis unless we can get the heavenly bodies to get their variable movements in line with our fixed systems.
Incidentally, measurement systems around the world, especially those allied to astronomy, are duodecimal. With our biases for ten and zero these days, twelve seems an awkward number, but read the following:
The duodecimal system (also known as base12 or dozenal) is a positional notation numeral system using twelve as its base. In this system, the number ten may be written as 'A' or 'X', and the number eleven as 'B' or 'E' (another common notation, introduced by Sir Isaac Pitman, is to use a rotated '2' for ten and a reversed '3' for eleven). The number twelve (that is, the number written as '12' in the base ten numerical system) is instead written as '10' in duodecimal (meaning "1 dozen and 0 units", instead of "1 ten and 0 units"), whereas the digit string '12' means "1 dozen and 2 units" (i.e. the same number that in decimal is written as '14'). Similarly, in duodecimal '100' means "1 gross", '1000' means "1 great gross", and '0.1' means "1 twelfth" (instead of their decimal meanings "1 hundred", "1 thousand", and "1 tenth").
The number twelve, a highly composite number, is the smallest number with four nontrivial factors (2, 3, 4, 6), and the smallest to include as factors all four numbers (1 to 4) within the subitizing range. As a result of this increased factorability of the radix and its divisibility by a wide range of the most elemental numbers (whereas ten has only two nontrivial factors: 2 and 5, with neither 3 nor 4), duodecimal representations fit more easily than decimal ones into many common patterns, as evidenced by the higher regularity observable in the duodecimal multiplication table. Of its factors, 2 and 3 are prime, which means the reciprocals of all 3smooth numbers (such as 2, 3, 4, 6, 8, 9...) have a terminating representation in duodecimal. In particular, the five most elementary fractions (½, ⅓, ⅔, ¼ and ¾), all have a short terminating representation in duodecimal (0.6, 0.4, 0.8, 0.3 and 0.9, respectively), and twelve is the smallest radix with this feature (since it is the least common multiple of 3 and 4). This all makes it a more convenient number system for computing fractions than most other number systems in common use, such as the decimal, vigesimal, binary, octal and hexadecimal systems, although the sexagesimal system (where the reciprocals of all 5smooth numbers terminate) does better in this respect (but at the cost of an unwieldily large multiplication table).

Decimal metric is the only way to go for distance measurements and money, don't get me wrong. But for calendars, spare me a dime!