Water words vary -
depending on the LOCATION  where water is found
(lakes, rivers, oceans, seas, wells, artesian springs, brooks, ponds, wadis, ocean/substance of creation) 
- or -
the FORM of water (dew, rain, snow, ice, glaciers, slush, sleet, thaw),
or its SUBSTANCE (fluid, liquid), or its USAGE (washing, pouring, drinking),
or its human EMANATION (urine, tears),
or  its form and extent of MOVEMENT (waves, froth, rapids, falls, floods, leaks). There are also other related "fluid" substances such blood, juice, semen, "milk" (mamma-ries) or "magma".

The modern-day linguists, incredibly  - in the days of clear tapwater - assume that water is simply H2O. In early days, it depended on where the fluid came from. Depending on the language group and environment, one of these forms came to be the dominant form for any language's "water"-word.

For example, if we take Latin "aqua" we find Latin AQUA "water" and Latvian AKA, ACINA dim. "well" Hebrew AJIN "spring, source", especially in arid regions, and Northwest Semitic 'AG(I)MU "lake?, troubled pool" (Sivan) i.e. a-gi-mu - frequently occurring in the texts whereas mema (me-e-ma) is rare (in later Hebrew this is JAM "ocean").

How can the linguists trace Latin AQUA to a non-existing hypothetical Indo-European root *akwa - for "water" where a wider non-blindered view of the evidence indicates that this particular "word" formation for water may have applied originally to a "well" or "artesian spring or source", i.e. a source of water under the ground - and not originally meaning "water" per se. Whether we can assign Northwest Semitic 'AGIMU and JAM to the above category can be disputed.

The linguists write - IMO correctly - that MAJIM i.e *ma-jim<*m-m
is at first glance a "dual" or more likely a plurale tantum,
especially since we also find the forms majja and majjin
in Aramaic and Syrian ("aqua, semen virile"), Akkadian mu, poetic Babylonian mam/wu and mj, Ethiopian maj, PLURAL mjm in Ugaritic and Egyptian mw (mjw) "water", mwj "urine, semen" - whence its metaphorical use for the concept of "son "and found also as mwj.t "dampness".
Compare here Latvian miz- or mist- (both long i) meaning "to take a leak".

I presume myself a "dual, plurale tantum" connection to the two human breasts of a woman and the fluid milk "of creation" for man's children, which we find as Indo-European *ma or plural and perhaps original onomopoetic Latin mamma "breasts", givers of fluid, which in Latvian are called pupas - and indeed, the Northwest Semitic texts, according to Sivan, show this consonantal shift between P > M.

Based on Sivan´s findings in Northwest Semitic texts, M- "water forms" are related to P- "water forms", whence Latvian UPE "river" > an M-form. Words such as Latvian MAZGAJAM "to wash" may also be related to MAJIM.

We also find Hebrew matar, Ugaritic MTR  "rain", where the -tar element is probably related to Hebrew TAL-, dew and anything falling from heaven, including drizzle, Latvian shal(t)- for which reason we have the following comparability of Hebrew matar "rain" with Latvian mitra "damp".
I doubt the relation of Hebrew matar "rain" to English "water" in spite of the possible relation of matar to water through m to w shift (see Bablyonian example of mam//wu above). Rather Latvian uden- udri- < *ten, *tro and English water, i.e. root *wed seem to relate to Latvian VADI or VADINI "veins, streams" and this is why many of the rivers of Europe are similarly named (TAN, DONAU, DNIEPER, ODER, etc.)

Latvian JURA "sea" (prononounced JOORA) finds its comparable in Hebrew JRH or JORAEH (variant MORAEH or MRH) - found as JE 'OR as the Hebrew+ name for the Nile but also JR "rain, raindrop" - to which NAHAR and NAHAL "wadis, Mesopotamia" are related, since, according to Sivan, Northwest Semitic shows the shift of M to N.

Hebrew SALAEG is the same as Latvian SNIEG both meaning "snow". Savin once again writes of the evidence for shifts of N to L in Northwest Semitic, weak, but nevertheless there, the change being in Semitic by some alleged to be through Hittite influence (which I doubt).

Hebrew SHATAH "to drink" is certainly related to Latvian DZE(R)T or DZE(R)TI "to drink", especially in view of Sivan's discussion of the evidence for shifts from DZ and TS forms to S and SH forms in Northwest Semitic.

Hebrew ZAERAM "rain" finds Latvian DZIRA, DZIRAM "swill, slop".

Even the Hebrew word GAESHAM (GSHM) "downpour of rain" is the same as Latvian GAZHAM pronounced GAHZHAM meaning "hefty downfall" - perhaps also related in some way to Latvian GAISS, GAISSAM "of the air" (whence GIZEH "to the air, upwards, the pyramids").

When we see that the original root of majim is simply MA- "fluid, water" and Latvian PLUdi, PLUO- is "flood" then Hebrew MABBUL as the term for the "great flood" is understandable. Indeed, water as a mythological and cosmic power is found as LIWJATAN (Leviathan) which is Latvian LIETIN- "rain".

The Latvian even has comparables to JORAEH "early rain" and MALQOSH "late rain in an agrarian seasonal context, i.e. Latvain AGRAJIE "early" and VELAKOSHIE "late".

The comparability of Hebrew and Indo-European is immense.