Readers may remember a couple of weeks back when I lamented that the downward economy had finally caught up to me and my little weather business. I asked readers if they’d mind if I’d showcase a weather station or two to help drum up some business. Nobody objected and many readers were very helpful and sympathetic. Thank you sincerely.
I always like to find ways to bring meteorological and climate science to the layman, hence this blog. My specialty is instrumentation, and readers may recall back in fall 2008 when I rigged up my vehicle with an NIST calibrated data logger, window mounted Gill IR shield, and a live connection to my laptop and drove a transect to measure the surface temperature. The result was this first ever graph of Reno’s UHI (Urban Heat Island) signature:
Click for larger image
That little experiment garnered quite a few comments, and I always wanted to find an inexpensive way for others to recreate what I did without having to invest in a high priced NIST calibrated probe, IR shield, datalogger, and laptop, all told which would run about $1200. That’s beyond the means of many people who might take an interest. It’s also well beyond the means of most students who might want to do something like this for their local science fair in their own hometown.
I’m pleased to say I have devised a solution for about 1/10th the cost, and I’ve just finished testing it in my hometown of Chico, CA. The results follow.
As many readers know, I sell some nifty USB dataloggers, and you may have seen this ad on the right sidebar which links to my online store.
The dataloggers are inexpensive, mostly under $100, and work great for various monitoring applications where you want to gather and evaluate temperature data. Problem is, they had no way to be mounted to anything easily, and I couldn’t see an easy way to mount one to a vehicle like I did with the IR Gill shield:
The NIST calibrated temperature sensor (inside the Gill IR shield) mounted on the vehicle using an improvised window mount., Cable goes to datalogger inside, and datalogger connected to laptop.
My goal has always been to recreate this inexpensively and reliably so many of you could map UHI in your own city. Making this within reach of students and their parents was the ultimate goal. How does $84.95 sound compared to $1200? You don’t even need to take a laptop in the vehicle with you.
Some days, serendipity strikes. A discussion about US flags with fellow blogger Russ Steele made me think about this idea of using the window flag mounts commonly seen en-route to football and baseball games:
Except my version has no team flag, it flies a datalogger instead:
How’s that for simple?
I spent today working this product idea out and testing it to make sure it is simple enough for anyone to recreate anywhere. Then I ran some road tests. Here’s how it works.
First we start with the USB-2 Temperature and Humidity datalogger. The reason I chose this unit over a temperature-only USB-1 version is that it already has a built in mini IR shield that allows free airflow. It reacts faster for that reason.
Next I provide a modified vehicle window mount and mounting hardware specifically to keep the USB-2 datalogger affixed, click images to enlarge:
Adhesive Velcro strip plus a removable secure strap worked best:
I ran some tests to see if I could fling the datalogger off the end. It held even after inflicting a few G’s on it at arms length.
Here’s the finished product:
The next test was to see if the datalogger was far enough out of the vehicle slipstream to prevent any temperature contamination while moving. That was evident from the data right away.
I did four daytime driving transects of my town to see if there were any hiccups or gotchas with install, driving, removal, and data download. Finding none, I wrote up some instructions:
Measuring UHI in your city:
- Install the software on your Windows PC (or laptop if you bring it with you in the car). Experiment with and familiarize yourself with the datalogger.
- Choose a route you can drive in your town that provides a good north-south and/or east-west driving transect from country, through suburbs, downtown, and back to countryside again. Use Google Earth or a similar mapping tool to assist in planning your route. Try to choose a route that will not take any more than 30 minutes to minimize ambient temperature change that may occur.
- Choose a clear night with no wind or light wind if possible. Later at night like 11PM-midnight is best so as to have minimum traffic and stoplights.
- Set the data logger to start logging either immediately or at some start time in the future. Put the waterproof cap back on
- Affix the datalogger to the vehicle mount using the strap provided.
- Put the vehicle mount on the top edge of your vehicle window and close it for a tight fit.
- Drive to the start point, Either wait for the pre-programmed start time or trigger the data logger to start from your laptop. It is always best to start on an easy to check period like top of the hour, quarter hour, or half hour.
- Drive the route, trying to maintain a fixed speed.
- Have your passenger note landmarks/locations in a logbook as you drive so that you can correlate with temperatures later.
- Optional – take a logging GPS unit with you if you want exact positions and time.
- Once completing the first transect, turn around and drive it again. You may wish to do several transects if time allows so that you can create an average profile later.
- Remove the USB-2 datalogger from the vehicle mount, plug into your PC, launch the included software and plot your results! Export the data to a spreadsheet for further analysis.
Saturday night, April 3rd, I ran my first for real UHI transect on Highway 99 from Garner Lane North of town to Neal Road South of town. I chose these points because they were safe places to turnaround and because they were both well enough removed from the population center that I expected they would approximate “rural” conditions.
After driving a transect and reverse driving it again, here is what the datalogger showed from the program that ships with the datalogger:
As you can see, the red temperature curve shows two humps, these are the two transects mirrored. First one was driving NW to SE, and the reverse (second) was SE to NW. Humidity and dewpoint are also displayed. The temperature looks a little blocky becuase the resolution of the datalogger is 0.5°C. However, given the world’s surface data is measured to the nearest degree, this is more than adequate for the purpose.
My home weather station (in town) showed light winds less than 5mph from the southeast and 50°F (10°C) when I left. Skies were clear. It was a good evening to measure UHI.
I took a voice recorder with me to note the time I started driving and the time I turned around and reversed the transect. I drove the reverse transect about 10mph faster because the first transect I didn’t want to pass a semi truck (speed limit 55mph for trucks through town) and pick up any waste heat, so I held well behind it.
Next I exported that data into my favorite graphing program ( DPlot ) and edited the starting and ending data times based on my recorder notes. I had set the datalogger at home to automatically start at 8:45 PM and I hadn’t reached my Garner lane start point yet. My actual transect began at 8:47 PM from Garner Lane at Highway 99 intersection.
Then I pulled up my Google Earth and started exporting some images for reference. Here’s the normal map view of the transect with starting and ending points:
And here’s the NW to SE driving transect overlaid onto a Google Earth 3D image looking NE so that Highway 99 appears as a “x- axis”:
As you can see, the graph matches the infrastructure character of the town fairly well, with the peak coinciding with the center of town (marked by the crosshair and label). The temperature is less in the rural outskirts of town.
So there we have it, it appears that my small college town of ~ 80,000 people has a UHI value of about 1.5°C (2.7°F) via this transect. The second return transect from SE to NW has the same magnitude, but is offset due to ambient temperature drop that occurred during the drive. When I returned home at 9:30PM, and hour later my home weather station had dropped to 44°F.
So would you like to do this for your hometown? Do you have a son/daughter nephew/niece that would benefit from a weekend science project learning about how to measure temperature and graph it to show UHI? Or, maybe you just want to do it yourself for fun and quantify what you’ve noticed for years via your automobile thermometer.
I’ll be happy to post any results here.
The only real caveat to this experiment is: don’t stop your vehicle if you can help it. Without a constant airstream, some waste heat from the vehicle might make its way to the sensor. Choose your driving route to avoid long stoplights, and I’ve found that in many cities after 11PM certain thoroughfares are “greenlighted”.
If you like this little idea or know somebody who would, please feel free to buy my little UHI measurement kit at a price most anyone can afford: $84.95 It helps me and provides more data that shows UHI to be a real, measurable, effect.
Here is the link to purchase the kit: http://weathershop.com/usb2-UHI.htm
And yes, I do ship internationally.
Thanks for your consideration, happy motoring! – Anthony
UPDATE: Hu McCulloch reminds me of the similar experiment Warren Meyer and his son did a couple of years ago which you can read about here:
http://www.climate-skeptic.com/2008/02/measureing-the.html
Some commenters were concerned about contamination of the data due to the vehicle. This is possible, even likely, at a dead stop. But at highway speeds of 60 mph or even street speeds of 30 mph any such contamination gets swept away in the slipstream. Essentially the thermometer has turbocharged aspiration, and is measuring the true temperature of the air being intersected. There may be issues with wast heat from vehicles ahead, but if you drive carefully and avoid tailgating, this can be avoided. Doing multiple transsects over time and averagign the runs is another way to minimize such biases.
========================================
The raw data from the USB datalogger is below, note the date format is dd/mm/yyyy.
ChicoUHILogger,Time,Celsius(°C),Humidity(%rh),dew point(°C),Serial Number 1,03/04/2010 20:45:00,9.5,58.5,1.8,010004039 2,03/04/2010 20:45:10,9.5,59.0,1.9 3,03/04/2010 20:45:20,9.5,59.5,2.0 4,03/04/2010 20:45:30,9.5,59.5,2.0 5,03/04/2010 20:45:40,9.0,60.5,1.8 6,03/04/2010 20:45:50,9.0,60.5,1.8 7,03/04/2010 20:46:00,9.0,61.0,1.9 8,03/04/2010 20:46:10,9.0,60.5,1.8 9,03/04/2010 20:46:20,9.0,61.0,1.9 10,03/04/2010 20:46:30,9.0,61.0,1.9 11,03/04/2010 20:46:40,9.0,61.5,2.0 12,03/04/2010 20:46:50,9.0,61.5,2.0 13,03/04/2010 20:47:00,8.5,62.5,1.8 14,03/04/2010 20:47:10,8.5,63.0,1.9 15,03/04/2010 20:47:20,8.5,63.0,1.9 16,03/04/2010 20:47:30,8.5,62.5,1.8 17,03/04/2010 20:47:40,8.5,62.0,1.6 18,03/04/2010 20:47:50,8.5,60.5,1.3 19,03/04/2010 20:48:00,8.5,60.5,1.3 20,03/04/2010 20:48:10,8.5,60.5,1.3 21,03/04/2010 20:48:20,9.0,60.0,1.7 22,03/04/2010 20:48:30,9.0,59.5,1.5 23,03/04/2010 20:48:40,9.0,59.0,1.4 24,03/04/2010 20:48:50,9.0,58.5,1.3 25,03/04/2010 20:49:00,9.0,58.0,1.2 26,03/04/2010 20:49:10,9.0,58.0,1.2 27,03/04/2010 20:49:20,9.5,57.5,1.5 28,03/04/2010 20:49:30,9.5,57.0,1.4 29,03/04/2010 20:49:40,9.5,57.0,1.4 30,03/04/2010 20:49:50,9.5,56.5,1.3 31,03/04/2010 20:50:00,9.5,56.5,1.3 32,03/04/2010 20:50:10,9.5,56.5,1.3 33,03/04/2010 20:50:20,10.0,56.5,1.8 34,03/04/2010 20:50:30,10.0,56.5,1.8 35,03/04/2010 20:50:40,10.0,56.0,1.6 36,03/04/2010 20:50:50,10.0,55.5,1.5 37,03/04/2010 20:51:00,10.0,55.5,1.5 38,03/04/2010 20:51:10,10.0,55.5,1.5 39,03/04/2010 20:51:20,10.0,55.5,1.5 40,03/04/2010 20:51:30,10.0,55.5,1.5 41,03/04/2010 20:51:40,10.0,55.5,1.5 42,03/04/2010 20:51:50,10.0,56.0,1.6 43,03/04/2010 20:52:00,10.0,56.5,1.8 44,03/04/2010 20:52:10,10.0,56.0,1.6 45,03/04/2010 20:52:20,10.0,56.5,1.8 46,03/04/2010 20:52:30,10.0,56.0,1.6 47,03/04/2010 20:52:40,10.0,55.5,1.5 48,03/04/2010 20:52:50,10.0,55.0,1.4 49,03/04/2010 20:53:00,10.0,54.5,1.3 50,03/04/2010 20:53:10,10.0,54.5,1.3 51,03/04/2010 20:53:20,10.0,54.5,1.3 52,03/04/2010 20:53:30,10.0,55.0,1.4 53,03/04/2010 20:53:40,10.0,54.5,1.3 54,03/04/2010 20:53:50,10.0,54.5,1.3 55,03/04/2010 20:54:00,10.0,54.5,1.3 56,03/04/2010 20:54:10,10.0,55.0,1.4 57,03/04/2010 20:54:20,10.0,55.5,1.5 58,03/04/2010 20:54:30,10.0,55.5,1.5 59,03/04/2010 20:54:40,10.0,57.0,1.9 60,03/04/2010 20:54:50,10.0,57.0,1.9 61,03/04/2010 20:55:00,10.0,57.5,2.0 62,03/04/2010 20:55:10,9.5,58.0,1.7 63,03/04/2010 20:55:20,9.5,58.0,1.7 64,03/04/2010 20:55:30,9.5,58.0,1.7 65,03/04/2010 20:55:40,9.5,57.5,1.5 66,03/04/2010 20:55:50,9.5,57.5,1.5 67,03/04/2010 20:56:00,9.5,58.5,1.8 68,03/04/2010 20:56:10,9.5,58.5,1.8 69,03/04/2010 20:56:20,9.5,59.5,2.0 70,03/04/2010 20:56:30,9.5,60.0,2.1 71,03/04/2010 20:56:40,9.0,60.5,1.8 72,03/04/2010 20:56:50,9.0,60.5,1.8 73,03/04/2010 20:57:00,9.0,61.0,1.9 74,03/04/2010 20:57:10,9.0,61.0,1.9 75,03/04/2010 20:57:20,9.0,61.0,1.9 76,03/04/2010 20:57:30,9.0,61.5,2.0 77,03/04/2010 20:57:40,9.0,61.5,2.0 78,03/04/2010 20:57:50,9.0,61.5,2.0 79,03/04/2010 20:58:00,9.0,61.0,1.9 80,03/04/2010 20:58:10,9.0,61.0,1.9 81,03/04/2010 20:58:20,9.0,61.0,1.9 82,03/04/2010 20:58:30,9.0,61.0,1.9 83,03/04/2010 20:58:40,9.0,60.5,1.8 84,03/04/2010 20:58:50,9.0,61.0,1.9 85,03/04/2010 20:59:00,9.0,60.5,1.8 86,03/04/2010 20:59:10,9.0,60.5,1.8 87,03/04/2010 20:59:20,8.5,61.0,1.4 88,03/04/2010 20:59:30,8.5,62.0,1.6 89,03/04/2010 20:59:40,8.5,62.5,1.8 90,03/04/2010 20:59:50,8.5,62.5,1.8 91,03/04/2010 21:00:00,8.5,62.5,1.8 92,03/04/2010 21:00:10,8.5,62.5,1.8 93,03/04/2010 21:00:20,8.5,63.0,1.9 94,03/04/2010 21:00:30,8.5,63.0,1.9 95,03/04/2010 21:00:40,8.5,64.0,2.1 96,03/04/2010 21:00:50,8.5,64.5,2.2 97,03/04/2010 21:01:00,8.5,64.0,2.1 98,03/04/2010 21:01:10,8.5,64.0,2.1 99,03/04/2010 21:01:20,8.5,64.0,2.1 100,03/04/2010 21:01:30,8.0,64.5,1.7 101,03/04/2010 21:01:40,8.0,65.0,1.8 102,03/04/2010 21:01:50,8.0,65.5,1.9 103,03/04/2010 21:02:00,8.0,66.0,2.0 104,03/04/2010 21:02:10,8.0,66.0,2.0 105,03/04/2010 21:02:20,8.0,66.5,2.2 106,03/04/2010 21:02:30,8.0,65.5,1.9 107,03/04/2010 21:02:40,8.0,64.5,1.7 108,03/04/2010 21:02:50,8.0,65.0,1.8 109,03/04/2010 21:03:00,8.0,66.0,2.0 110,03/04/2010 21:03:10,8.0,66.0,2.0 111,03/04/2010 21:03:20,8.0,66.0,2.0 112,03/04/2010 21:03:30,8.0,66.0,2.0 113,03/04/2010 21:03:40,8.0,65.0,1.8 114,03/04/2010 21:03:50,8.0,64.0,1.6 115,03/04/2010 21:04:00,8.0,63.5,1.5 116,03/04/2010 21:04:10,8.0,63.0,1.4 117,03/04/2010 21:04:20,8.0,62.0,1.2 118,03/04/2010 21:04:30,8.5,61.0,1.4 119,03/04/2010 21:04:40,8.5,61.5,1.5 120,03/04/2010 21:04:50,8.5,61.0,1.4 121,03/04/2010 21:05:00,8.5,60.5,1.3 122,03/04/2010 21:05:10,8.5,60.5,1.3 123,03/04/2010 21:05:20,8.5,60.0,1.2 124,03/04/2010 21:05:30,9.0,60.5,1.8 125,03/04/2010 21:05:40,9.0,61.5,2.0 126,03/04/2010 21:05:50,9.0,61.0,1.9 127,03/04/2010 21:06:00,9.0,61.0,1.9 128,03/04/2010 21:06:10,9.0,61.0,1.9 129,03/04/2010 21:06:20,9.0,61.0,1.9 130,03/04/2010 21:06:30,9.0,60.5,1.8 131,03/04/2010 21:06:40,9.0,60.0,1.7 132,03/04/2010 21:06:50,9.0,59.5,1.5 133,03/04/2010 21:07:00,9.0,59.0,1.4 134,03/04/2010 21:07:10,9.0,59.0,1.4 135,03/04/2010 21:07:20,9.0,59.0,1.4 136,03/04/2010 21:07:30,9.0,58.5,1.3 137,03/04/2010 21:07:40,9.0,58.5,1.3 138,03/04/2010 21:07:50,9.0,58.5,1.3 139,03/04/2010 21:08:00,9.0,58.0,1.2 140,03/04/2010 21:08:10,9.5,58.0,1.7 141,03/04/2010 21:08:20,9.5,57.5,1.5 142,03/04/2010 21:08:30,9.5,57.5,1.5 143,03/04/2010 21:08:40,9.5,57.0,1.4 144,03/04/2010 21:08:50,9.5,57.0,1.4 145,03/04/2010 21:09:00,9.5,57.0,1.4 146,03/04/2010 21:09:10,9.5,57.0,1.4 147,03/04/2010 21:09:20,9.5,57.0,1.4 148,03/04/2010 21:09:30,9.5,57.0,1.4 149,03/04/2010 21:09:40,9.5,57.0,1.4 150,03/04/2010 21:09:50,9.5,57.0,1.4 151,03/04/2010 21:10:00,9.5,56.5,1.3 152,03/04/2010 21:10:10,9.5,57.0,1.4 153,03/04/2010 21:10:20,9.5,57.0,1.4 154,03/04/2010 21:10:30,9.5,57.0,1.4 155,03/04/2010 21:10:40,9.5,59.0,1.9 156,03/04/2010 21:10:50,9.5,61.0,2.4 157,03/04/2010 21:11:00,9.5,61.5,2.5 158,03/04/2010 21:11:10,9.0,61.0,1.9 159,03/04/2010 21:11:20,9.0,61.0,1.9 160,03/04/2010 21:11:30,9.0,61.0,1.9 161,03/04/2010 21:11:40,9.0,62.0,2.1 162,03/04/2010 21:11:50,8.5,62.0,1.6 163,03/04/2010 21:12:00,8.5,62.5,1.8 164,03/04/2010 21:12:10,8.5,63.0,1.9 165,03/04/2010 21:12:20,8.5,63.5,2.0 166,03/04/2010 21:12:30,8.5,64.0,2.1 167,03/04/2010 21:12:40,8.5,63.5,2.0 168,03/04/2010 21:12:50,8.0,63.5,1.5 169,03/04/2010 21:13:00,8.0,64.0,1.6 170,03/04/2010 21:13:10,8.0,64.0,1.6 171,03/04/2010 21:13:20,8.0,64.5,1.7 172,03/04/2010 21:13:30,8.0,65.5,1.9 173,03/04/2010 21:13:40,8.0,65.0,1.8 174,03/04/2010 21:13:50,8.0,64.0,1.6 175,03/04/2010 21:14:00,8.0,63.5,1.5 176,03/04/2010 21:14:10,8.0,63.0,1.4 177,03/04/2010 21:14:20,8.0,62.5,1.3
What’s this about surface temperature data having raw resolution of only 1.0 DegreeC?
How much averaging does it take to make it meaningful when the magnitude of trends being looked for is small?
My rule of thumb would be 0.1 degreeC, recognizing that is data resolution not accuracy. 0.5 is OK for your homespun proposal, which is a great idea for schools, but shouldn’t official sensors be far better?
I was thinking about IR bias from the heat source you bring with you. I’m not sure of the absorptivity of the datalogger itself, but you could shield with a small disk of material below, with a small amount of gold coated tape on the bottom. Gold has a great reflectivity of IR. Gold tape is expensive, but in this small an amount, it shouldn’t add terribly to the cost.
Just a thought.
why not mount such a system on all public vehicles, such as buses? Taxis? commuter trains? These run all the time. Certainly, with enough data points we could at least get trends from in city to suburbs, regardless of factors pulling us one way or the other.
Asking the city or any government agency to do this (on public vehicles, for example) turns the project back to the folks who brought you GISSTEMP and CRU. Much better for us amateurs to do it. Sure, it will have some errors, but with enough data from enough places, we’ll get a good picture. Driving around airports multiple times may get TSA a bit upset, but freeways through cities should be easy.
Any thoughts on how and where to send or collect results? Formats? I’m OK with spreadsheets, but Windows code is not my strong point. (If you need code for an embedded 8 bit micropower processor, give me a call.)
Hey, I just did my first run and it was such a blast and I’m so pumped up I can’t even believe how fun this is! I am so excited I haven’t read everyone’s comments yet but reading the first data set reminds me of the 0th law of thermodyamics – that is – let the transducer stabilize to the outside temperature before you start the log. (duh…on me). Anyway, I live in the desert southwest right close to the open desert and I got a phenomenal UHI just based on coming in off the desert and back into town. I will not post my results yet as I want to at least run five or so samples just to make sure my set up is right but you will be hearing from me.
OT – is it just me or are “skeptics” having a lot more fun than the AGW crowd?
REPLY: We look forward to it. Thanks, – Anthony
Pity the software is Windows. I run Debian/GNU-Linux on more than half a dozen desk top machines and three laptops. How about porting the software to the GNU-Linux operating system?
Received the datalogger yesterday and tested it for 1 hour (on my desktop) last night. After downloading the data to my PC (and plotting it), I reset the datalogger for a 24 hour run, and placed it outside.
Note: I used the °F setting and noticed that its resolution is 1°F, vs 0.5°C (obvious).
Since I can get to about 100 yards from the San Antonio International Airport temp sensor, I’ll use that point to start my runs. (The SAIA sensor is at the end of a runway, which is about 100 yards from a security fence that runs along a freeway access road – not too busy even in the daytime).
UHI question: Since UHI is larger at night, how long after sunset would show the max value? Would be interesting to check delta UHI at different times.