MFTC Modified Cold Hardiness MapBETA

We find the USDA hardiness map limited in a few important ways.

It shows you the average coldest night of the year, averaged out over thirty years. But a plant doesn't die to an average. It dies to the one bad night in thirty. Across the country that gap runs 15 to 34°F, which is two to three whole zones, and it always runs the wrong direction. Your USDA zone makes your winters look easier than they really are.

The other problem is that it's just plain hard to match a plant up to a zone number. Zones are broad labels, and a label doesn't tell you as much as a temperature does.

We rate every plant by the lowest temperature it's actually been seen to survive. Balaton tart cherry is probably good to about −35°F. Evan's Bali will take −40 to −45°F. That's a number you can do something with.

That comes from real winters, not from a model. Mostly from our own two growing sites in Missoula and Florence, Montana, and from fruit growing friends up north that we trust. The late Rod McIver in Kalispell, Bob Purvis in Homedale, Tom McCamant at Forbidden Fruit in Paradise, Ed Schultz in Gallatin Gateway, Steve Masterman up in Fairbanks, Bernie Nikolai in Edmonton. Plus years of reports from customers. Where we don't have a local record we lean on published nursery and university numbers.

We're not the only ones doing it this way. Bob Purvis, Steve Masterman of Alaska Fruit Trees, and others rate hardiness right in line with how we do it. Plenty of folks out there don't, probably because their numbers aren't based on what the plant has survived down to. That's usually why you'll see a variety listed warmer somewhere else than we list it.

1. We find your address on the map. It gets turned into a latitude and longitude. Nothing gets stored and nothing gets sent to us.
   Source: US Census Bureau geocoder
2. We pull the long record weather stations near you. Usually the closest eight, each one with at least ten winters on file, often thirty or more. For each one we grab the single coldest night of every winter going back to 1995.
   Source: NOAA Regional Climate Centers (ACIS). NWS co-op, airport, SNOTEL and RAWS networks
3. We work out the 1 in 30 year low for each station. Those yearly coldest nights get run through an extreme value fit. It's the standard way of doing return periods, same math behind a hundred year flood.
   Method: Gumbel extreme value fit to annual minima
4. We weight the stations by how much they look like your site. Closer counts for more, and so does matching elevation and a long record. A station 500 feet below you sitting next to a lake sees different winters than you do, so it counts for less. Then we take an average that a couple of oddball stations can't drag around.
   Method: distance, elevation and record length weighting, Huber robust centre
5. We read the lay of your land. Off a 100 foot resolution elevation model we measure how much of the ground around you sits lower, how far cold air has to travel before the ground drops 50 feet, and how closed in your horizon is. Cold air is heavy. It runs downhill and puddles up wherever it can't get out.
   Source: USGS / Mapzen elevation data
6. You tell us what the stations can't know. That's question 2 up at the top. There's no public data anywhere that can fill it in for you.
   Source: you
7. We match that up against what our plants have survived. Every variety we grow carries a temperature it's been seen to come through. We show you the ones that clear your number.
   Source: our own records and our grower network

Tested against about 1,650 weather stations we held back, the estimate lands within 3.5°F half the time. We also hold the extreme to be colder than the average no matter what, because it can't come out warmer. That isn't physically possible.

We're a mission oriented company. Our mission is to increase food production capacity for localities in Montana and beyond, and we do that by inspiring folks to grow fruit, nut, and berry producing trees, shrubs, and vines.

This map isn't perfect and we're not going to pretend it is. But for what we're after, we think it's better than what's out there right now. If somebody plants a tree and it lives, they'll plant another one.