This time of year the heart of the Arctic is frozen solid, and the only changes occur in the marginal seas.  Above shows the Atlantic basins, especially Kara, Barents, Greenland Sea and Baffin Bay.  All of them seesawed during the month, with some fall off at the end, especially noticeable in Gulf of St. Lawrence (counted with Baffin Bay).

Meanwhile on the Pacific side, Bering fluctuated, while Okhotsk lost extent steadily toward month end.

Context

The monthly ice extent average for March provides indication of any year’s annual maximum, prior to melting down to the September annual minimum. Sometimes a lower March extent yields a lower September extent, but not always: 2012 had both the highest maximum and lowest minimum in the last 11 years. That was the year of the Great Arctic Cyclone, and an outlier in the record.

Looking at the 11-year averages in the MASIE data set, the pattern in round numbers is:
Maximum: 15.0 M km2
Minimum: 4.8 M km2
Loss: 10.2 M km2
Loss: 68.0 % of maximum

So about 2/3 of the maximum extent is lost, varying from 66 to 70%. Obviously, all the factors affecting ice extents are in play: (Water, Wind and Weather) with the September outcome uncertain, but likely to be in the range observed.

March 2017 in Comparison

As has been reported, ice formation this year has been sluggish compared to other years. The graph below shows March 2017 compared with the 11 year average, and with 2006 and 2016, as well as SII (Sea Ice Index).

This March started below average, lost sllghtly until the third week, then recovered some before dropping off at the end. 2006 dropped off more rapidly than 2017, while 2016 ended near average. SII showed lower extents all month but drew close at the end.

The Table below shows Day 90 extents across the Arctic Seas compared to averages and 2006, the lowest recent year.

Region	2017090	Day 090 Average	2017-Ave.	2006090	2017-2006
(0) Northern_Hemisphere	14228992	14791162	-562170	13913402	315590
(1) Beaufort_Sea	1070445	1070018	427	1068683	1762
(2) Chukchi_Sea	966006	965297	709	959091	6915
(3) East_Siberian_Sea	1086168	1085794	374	1084627	1541
(4) Laptev_Sea	897845	896573	1272	897773	71
(5) Kara_Sea	831189	924617	-93428	922164	-90974
(6) Barents_Sea	525362	656247	-130885	623912	-98550
(7) Greenland_Sea	705581	661500	44081	604935	100645
(8) Baffin_Bay_Gulf_of_St._Lawrence	1467334	1426694	40641	1026934	440401
(9) Canadian_Archipelago	853214	852652	562	851691	1523
(10) Hudson_Bay	1260903	1251383	9521	1240389	20514
(11) Central_Arctic	3247995	3235035	12960	3241074	6921
(12) Bering_Sea	702504	847340	-144836	662863	39640
(13) Baltic_Sea	29767	75051	-45284	129348	-99580
(14) Sea_of_Okhotsk	575084	830273	-255189	588167	-13083
(15) Yellow_Sea	0	99	-99	1067	-1067
(16) Cook_Inlet	7318	5460	1858	5462	1856

The marginal seas in the Atlantic and Pacific make the 2017 deficits to average: especially Barents, Kara, Bering and Okhotsk. Those seas usually lose all their ice by September. 2017 Surpluses in Greenland Sea and Baffin Bay are smaller, but make most of the difference with 2006.

2017 Outlook

March this year averaged 14.509 M Km2 compared to the 11 year average of 14.986 M km2, a deficit of 478k km2 or 3.2% down. That suggests that a typical melt later this year would result in a minimum of about 4.5 or 4.6 M km2, slightly down from the 11 year average of 4.8M km2.

Sea Ice Index (SII) typically shows less ice than MASIE, and SII reports a 2017 March average ice extent of 14.273 M km2 compared to SII 11 year March average of 14.842, a drop of 569k km2 or 3.8%.  Folks relying on SII may be expecting a lower September minimum, perhaps even breaking the present plateau of ice extents since 2007.  That remains to be seen.