Public Lab Research note


Zero dark thirsty: Capturing replenishing nighttime rain events with a trail camera

by cfastie , warren | September 20, 2016 03:53 20 Sep 03:53 | #13465 | #13465

Above: LEAFFESTers observing the deployment of a trail camera (arrow) near a small stream.

Jeff brought a new trail cam to LEAFFEST to see how it worked for monitoring streams. There are many potential uses of this type of camera for documenting disturbances to water bodies. The cameras are designed to be installed outside and take photos or video at regular intervals or when motion is detected. They are sensitive to near infrared (NIR) light and have NIR LEDs to illuminate the scene at night. In addition to capturing animal activity, they could record when unnatural discharges into streams occur. This could provide important evidence in many situations where human activity might be compromising stream health.

Crenova2.JPG
Above: The model of trail camera we installed by a stream during LEAFFEST. I don't know what all the things on the front of the camera do.

We strapped the camera to a tree after dark and pointed it upstream along Halnon Brook. The stream flow has been very low, and in the last month aboveground flow has been intermittent (it was completely dry for short stretches) for the first time in about 14 years.

Moto_20160908_151611174c.jpg
Above: this is the view of the stream on September 8, 2016, nine days before the trail camera was installed. The flow was intermittent.

We did not install a measuring pole in the scene to get good information about changes in water depth. We expected a half inch of rain overnight, but did not expect a dramatic surge in stream flow because the watershed is very dry, well forested, and completely undeveloped.

_IMAG1212_inset2k.jpg
Above: One of the photos taken by the trail camera in the morning after the overnight deployment. The inset is a ruler in front of a rock that provides a good gauge for water depth in the timelapse video made from the series of photos. The increments on the left side of the ruler are inches. The inset was taken on September 19 after additional rain fell and the water depth was four inches -- about as high as it was for the entire trail camera deployment.

The trail camera was set to take one photo every minute. When it is dark, the camera automatically turns on its NIR LEDs for each exposure and captures an infrared photo. When it got a little bright in the morning, it stopped using the LEDs but continued recording grayscale NIR photos (note the whitish leaves). When it got bright enough, it started taking regular color photos.

Between 9:22 PM Saturday night and 12:32 PM Sunday afternoon the camera captured 910 photos. These were added to Adobe Premiere Elements as "numbered stills" and became the frames of a 30 second video with 30 frames per second. The camera puts an information bar at the bottom of each photo with the date, time, and temperature. In addition, I added some text annotations, but the video has no audio.

The video records two periods of rain during the night, and each downpour results in a noticeable rise in stream water level of about an inch. Although the rain had stopped by 7:00 AM, at 11:10 AM the level of the stream quickly rose an inch or more. At the end of the video I repeat several times a three second clip which spans the time from 11 AM until noon. During this time you can see the quick water level rise, especially in a pool above the fallen log. This must have been a surge flowing downstream from a rain event high in the watershed. I don't know how much time elapsed between the rain event and the appearance of the surge near the camera, but it had not rained near the camera for more than four hours.

The quality of the color photos taken by the camera is very good. The quality of the grayscale NIR photos taken in good light is also good. The quality of the NIR photos taken under illumination by NIR LEDs is not so good. The photos are grainy because the LEDs do not provide much light and the sensor is being operated at a very high ISO. These NIR photos are also vignetted because the LEDs do not illuminate the edges of the scene well. There is still a lot of information in the photos, but the evidence of pollution events at night will be very different from daytime evidence. And nocturnal bears and wapiti captured at the edges of the frame will be very dark.

Jeff wants to post this activity as an "activity" in a format that others can follow so they can repeat what we did. I was going to do that, but old habits are hard to break. This might be a good opportunity to compare the two types of posts and see what the advantages of each are.


7 Comments

TY so much for posting the details of this test @cfastie and @warren!

Reply to this comment...


Oh no @Zengirl2, TY so much for posting the awesome hard-hitting journalistic expose of LEAFFEST at the Adafruit blog: https://blog.adafruit.com/2016/09/20/what-happens-when-you-camp-with-science-tech-lovers-publiclab/. It was great to meet you at LEAFFEST.

Reply to this comment...


This was fun and pretty easy. I'd love to try collaborating on a follow-up activity, I don't think it'd be too complicated. What do you think?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Don't we have to have an activity before we have a follow up activity? There are lots of ways to make this an activity and many ways to follow up (various cameras, different stream types, ruled post in the scene, different intervals, daylight only, video clips vs stills, motion detection, near unnatural stream inputs, etc.). This activity can be taken to many places. Someone should impose some directional constraints. Maybe I will try to replicate this activity with a Mobius ActionCam.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Sorry, I meant make an activity out of this, with step by step instructions and such, not one taking the next step. So, yes!

I'm hoping @stevie and the folks she's working with who asked us to do this test can provide some more specifics on where and how they're hoping to do this sort of thing, so we can put more details and constraints into such an activity. Anybody the things Chris mention jump out at you, Stevie?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Is the problem that the mobius can t do intervals greater than a certain amount? or mostly it's the NIR and motion capture that you are looking for?

S

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Scott,

There is no problem. The Mobius can do interval shooting at any interval, and it can also do motion detection. The trail camera has NIR LEDs to illuminate the scene for night time photography. These LEDs don't have to be NIR for stream monitoring -- the trail cam uses NIR so the wildlife doesn't see it and get scared. So any light source could be used with a Mobius for night time stream monitoring.

The trail cam is already weatherproof, so no enclosure is needed. The trail cam has smart electronics so it conserves power between shots and can last for months on one set of batteries. The Mobius can also do well with power in some situations.

Some trail cams don't cost much more than a Mobius, and they are ready out-of-the-box to deploy for weeks of 24/7 all-weather interval or motion detection shooting.

Chris

Reply to this comment...


Login to comment.