Best of PhenoCam

As I wander through the PhenoCam archives, I encounter truly stunning or interesting images. From time to time I’ll post these “best of PhenoCam” images on my blog together with some location information.

This first image is an old one, posted to my Flickr site a long time ago. It shows an early sunrise at a boreal Jack pine (Pinus banksiana) forest roughly 30 km south of Chibougamau, Quebec, Canada. Jack pines form fire adapted stands of 9-22 m tall and lanky trees. The image was taken from a 30 m high scaffolding tower harbouring an eddy covariance flux system.

[caption id=”” align=”alignnone” width=”640”] The sun peers through the morning fog illuminating the tops of a boreal Jack pine forest.[/caption]

Battery run

Past January, after 10 years of service, the battery bank of the Bartlett flux tower gave in to too many cycles and harsh working conditions. Since it is a key component of the flux tower, replacements were bought and a large part of the lab made a trip to Bartlett to carry 6 large deep cycle batteries into the New Hampshire woods. Below the rather funny video of me sledding a battery into the woods. After a day of work, things are back up and running and have been ever since, hopefully for another 10 years.

http://www.youtube.com/watch?v=l5RFjYbJ-hM

 

Dating the forest

Determining stand ages in a tropical humid forests remains a challenge, mainly due to the lack of (annual) growth rings in the wood of many tropical trees. A few months ago I was offered the opportunity to date wood core samples using radio carbon dating at the UC Irvine Keck Carbon AMS facilities, resolving the issue of a lack of visible characteristics to determine the age of a tree. I’m indebted to Prof. dr. Trumbore for providing this generous offer as these tests are expensive.

Three samples along a wood core for two tropical humid forest species were selected. One species consisted of a dominant canopy species Scorodophleus zenkeri and a common mid-sized / understory species Panda oleosa.

Only the  youngest of the three samples for each species was valuable in this exercise. Other samples fell within a region which is hard to date using radio carbon dating as the calibration curve for this region is relatively flat (~1700 - 1950) resulting in large uncertainties when relating 14C-age to calendar years (see examples Figure 1).

[caption id=”” align=”aligncenter” width=”600”] Figure 1. - Relationship between 14C-age and calendar years. Example 1 shows a clearly defined date for a given 14C-age. Example 2 shows multiple possible calendar years matching a given 14C-age, complicating accurate dating (courtesy of Utrecht University)[/caption]

 

Although the data is limited to two data points per species, the first being the time at which the wood core sample was extracted and a second the dated sample as determined by radio carbon dating, we could get a rough estimate of the growth rates for both species during the last few decades.

Ages of the most recent samples were determined to be 1962 / 1985 and 1966 translating into growth speeds of~0.8 / 1.3 mm/yr and 0.7 mm/yr for Scorodophleus zenkeri and Panda oleosa respectively. These values fall well within the ranges as defined by literature but on the lower end of the range.

The slow growth rate of Panda oleosa is not surprise as it is often light limited, being a mid-sized species. The stature of these understory trees pales in comparison to emergent trees such as Petersianthus macrocarpus (Figure 2), their age however might not. Take home message: don’t judge a tree’s age by it’s stature.

[caption id=”” align=”aligncenter” width=”339”] Figure 2 - Petersianthus macrocarpus, a potentially emergent tree reaching ~45m height with a diameter of ~150cm (courtesy http://liberianfaunaflora.org)[/caption]

Dell Optiplex 980 SMP (Ubuntu 12.04 LTS)

My Dell Optiplex 980 at work doesn’t seem to play nice with Ubuntu 12.04 LTS. It fails to detect hyper-treading support on the CPU and defaults back to a single core setup when effectively running a 8-core CPU. This is rather annoying.

It seems that acpi settings are to blame, adding:

acpi=ht

to your grub config file mediates this issue and enables SMP support. Sadly, adding this parameter is undone by kernel updates.

Octave 3.6.x image processing on Ubuntu Legacy

If you want to install the most recent version of Octave, an open source MatLab clone, on your older Ubuntu box you are out of luck. Only the most recent repositories (> 13.x) carry the latest Octave releases.

I found this quick fix for the problem. It involves four steps on a terminal.

sudo apt-add-repository ppa:octave/stable
sudo apt-get update
sudo apt-get install octave
sudo apt-get install liboctave-dev

To get to a fully functioning image processing toolbox run the following commands:

sudo octave

on the octave prompt

pkg -verbose install -forge general
pkg -verbose install -forge control
pkg -verbose install -forge specfun
pkg -verbose install -forge signal
pkg -verbose install -forge image

The command ‘pkg list’ will return a list of installed packages

pkg list
Package Name  | Version | Installation directory
--------------+---------+-----------------------
     control *|   2.6.0 | /usr/share/octave/packages/control-2.6.0
     general  |   1.3.2 | /usr/share/octave/packages/general-1.3.2
       image  |   2.0.0 | /usr/share/octave/packages/image-2.0.0
      signal  |   1.2.2 | /usr/share/octave/packages/signal-1.2.2
     specfun  |   1.1.0 | /usr/share/octave/packages/specfun-1.1.0

Finally, you load the image package using:

pkg load image

 

Pagination


© 2018. All rights reserved.

Powered by Hydejack v7.5.1