sudden eff drop (30 to 22MPG)

[TCMIV]

On the most recent fill-up, efficiency dropped to 22 MPG from the previously consistent 30 (+/- 1.5) MPG. I'd love to hear why you guys think this might be. Thanks!

Here's some relevant information followed by my hypotheses:

Model: all-stock 2009 Fit Sport AT w/o NAV
Odometer: 730
Purchased: 10/10/2008
Observed eff at refills (MPG):
(original tank filled by dealer)
1: 30 (refill at Grand Martco)
2: 31 (refill at ARCO AMPM)
3: 29 (refill at ARCO AMPM)
4: 28.5 (refill at Shell Menlo Atherton)
5: 22 (refill at Shell Menlo Atherton 11/17)

Narrative:
I've refilled several times since getting the car, refilling at around the halfway point on the fuel gauge.
Calculating from the trip meter (reset after each refill) and the amount the pump tells me it pumped until it clicks off, I've been getting pretty steady eff around 30MPG.
On refill 3 I went to ARCO AMPM. On refill 4 I went to Shell Menlo Atherton, obtaining about half a tank of gas from Shell Menlo Atherton. I observed 28.5 MPG on the gas previously obtained (from ARCO and Martco). Then today I went to Shell Menlo Atherton again to fill up, this time the gas gauge was only about 0.3 to 1/3 of the way from F to E. Trip meter was at 98.6. The pump says it dispensed 4.46ga until it clicked. So that works out to about 22MPG.

Hypotheses:
1. I drove it harder.
I don't feel that I drove it any different. And if anything, between fill 4 and fill 5, the duties appear to have been lighter because I carried passengers less often and climbed grades less often. The highway vs city mix also seems similar to previous fill intervals.

2. Tire pressure low.
This might be it. I started with 33psi on 10/10. And on 11/16 the tires were at 31psi. I took it as ordinary loss of pressure over the course of the month and added pressure today. So it makes sense that as the tires lost pressure efficiency went down. But the drop to 22 just seemed too sudden. Any thoughts here?

3. The pump overstated the amount of gas it pumped to full.
I was surprised to have read 4.46ga from the pump because the gas gauge reading led me to think I hadn't used up that much gas since the previous fillup. Of course, gas gauges are inaccurate, but that coupled with the suddenly lower efficiency calculated made me quite suspicious. Any thoughts on this?

4. Gas from Shell less efficient in my car.
Fill 4 was the first time I used Shell gas in this car, so on that level the theory seems plausible. Still, I find it hard to believe because of the great difference. Any thoughts on this?

5. I don't have another hypothesis. Other hypotheses would be appreciated!

[Hootie]

I've read somewhere on these boards that during colder season gas stations have been known to switch to fuels with more ethanol opposed to the straight gasoline they offer.

Also, the car could have been in VTEC more that usual. In my GD3 I was getting 26-30 MPG to and from school by shifting too early with the paddles which caused the VTEC to engage at around 2400 RPMS instead of 3400.

[CBX]

A third of a tankful on a Fit is not a very large 'sampling', where a half gallon variation at shut-off represents a significant percentage.

You might try the same pump at the same station when you are near empty, and compare results.

P.S. The low-fuel light does not start blinking when you are "dangerously" low!

[gimme]

fill up before last 39mpg, this fill up 32 and it's time for an oil change. i allways notice a drop in mpg when an oil change is around the corner. this is just my experience though

[TCMIV]

Thanks guys!

@Hootie:
AFAIK all the gas I get in California has Ethanol content in the 5%-10% band. It seems to be too narrow a band to make much difference?
And I'll look into the VTEC thing. I haven't read much about how iVTEC works or what it does to FE.

@CBX:
Yea I'm considering that or buying a 5ga gas can to test. Seems like too much work and expense to be justified. But running till low fuel indicator comes on sounds good. I just don't know whether the indicator comes on at a consistent enough fuel level or not. I'll have to do some digging.
Also I figured I'd get gas whenever it's convenient so I don't incur extra cost of making a dedicated trip to the station.

@gimme:
Thanks, that makes sense. Though I'm not anywhere close to an oil change right now. I'll keep that in mind as I approach the first service interval.

[CBX]

I wasn't intending to suggest using the fuel light as any measurement, but that you should use up more than three gallons prior to refueling and doing the consumption calculation.

If the pump shutoff point varies some on each fillup, that amount of variation will be a smaller percentage of nine or 10 gallons than of three.

[Texas Coyote]

Fill to the top on fill ups, try the other suggestions and if your mileage still sucks buy a locking gas cap. My performance and fuel mileage is always better on fresh Mobile1. I hope that you figure it out.

[TCMIV]

Thanks guys. I still don't know what happened exactly, but the anomaly never happened again (no matter what gas I used). I've been getting 31-33 MPG consistently since.

[nmfit2008]

Just to clarify, when the low gasoline light comes on, it should then take 9.4 gal. Since the capacity is 10.8, then you can still go at least 50 miles with the light on. However, over on the hypermiling thread, you will see that many people are able to put in at least 12 gallons in the tank, partly due to the long neck...as the tank is under the front seats. I don't recommend this to all.

I still recommend increasing your tire pressure to 40 psi, or even 45 psi, which i run in the Spring, Summer, Fall. Your FE will be so much better, and you pay no price whatsoever.

I live in very high mountains. Even though we do have winter, it is not as tough as winter in the east....so my mileage has dropped a bit, but I continue to maintain my over 40 average......that is a full 9 months of driving. Most of my driving is 'around town', with maybe 10% highway.