Friday, 20 November 2015

Some images from Nouragues: II

Our temporary office - 3 hammocks, lost of washing, all the kit, no walls.

Amazing view from above the camp on the Inselberg itself - a very strange-looking granite outcrop, with patchy scrub and exposed rock, among hills covered in green. Early morning mist out over the 100s of km2 of forest. 

Great example of the complex forest canopy structure and biochemistry. Or big broccoli as I like to see it.

Classic bit of dynamic geography in action.

Sorry. Not pretty but sums up the feeling. Slightly damp Blaise to the left and Any behind me.

Some images from Nouragues

Here are a few images from our time at the CNRS station at Inselberg, Nouragues.
The type of large, buttressed trees that hold all the carbon. How on earth do you model that trunk??

Obligatory lidar in the canopy shot.

And and Blaise working hard.

Perhaps the biggest hazard in the whole place: HORRIFIC spikes on the palm stems, under their leaves, everywhere. These things are an absolute nightmare. Shouldn't be allowed.

Scan 242. The last one. Slightly blurry, but then that's how we felt I think.

Mission complet

So we spent an amazing 10 days or so at the Inselberg field station, Nouragues. The weather was a bit mixed - the so-called dry season seemed to be ending a bit early, and so some days we had to flee from unbelievable rain storms. But working dawn until dusk, and following Andy's new Stakhanovite sampling regime of a 10m grid, with 10 rolling targets - 5 ahead, 5 behind - we completed 242 scans in the 1 ha H20 plot, at a level of detail we've never even got close to before. We were indebted to Blaise (Tymen) who was our 3rd man and guide, and he helped made the task manageable. So far, the data look absolutely stunning. Our first visualisation is here:

The location of the station is stunning - deep in the dense Amazonian jungle. The helicopter ride in and out was worth the trip in its own right, even if it was cloudy on the way in. It wasn't on the way out!

Notes on scanning and lessons learned?
Scanning at this density (4 x what we've done before), ditching the waveform data and the photos, we managed to achieve 40-50 scans per day, even with the odd shower, as opposed to our previous best of 23 previously. We could do this on 2 batteries, and without filling the disk on the scanner so only needing to download in the rapidly fading light of the evening, enabling us to leave the scanner at the plot (in the case in a waterproof bag and under a tarp!).

Doing it this way also meant we only ever needed 10 poles and targets (so you could take 20 say, to leave spares). We just needed to do a little bit of judicious veg shifting & pruning for each new location; we never missed one. Finally, it meant that laying out targets was very, very much quicker and simpler than using the systematic schemes we've used in a 20m spacing in the past. So, swings and roundabouts.

Andy was able to download and register the new data each day, to make sure we could continuously update our chain from the start to the end of the plot. The multi-station adjustment obviously takes exponentially longer as the number of scans grows, but we were able at least to check we could register all scans together.

Sunday, 8 November 2015

Off again! French Guiana here we come.

And so it's time to say goodbye to rainy England and head to hopefully not-so rainy French Guiana. We've been fortunate to get some funding via the French Government CNRS Nouragues Travel Grants Program, to visit the field site at Nouragues, and scan some existing tropical forests plots. Andy and I are ready to go, kit loaded on the plane and just hoping this isn't going to be too much of a problem:
Slightly foggy Orly airport.
We're looking forward to the helicopter ride to the forest site from Cayenne and getting to see a new patch of tropical forest. We're also excited to be able to visit some of the key tropical plots set up by Jerome Chave and colleagues. We know we're going to be seeing some lianas, and that ought to be interesting if challenging too.

UPDATE: arrived safe and sound in Cayenne, with all kit which is the important thing. Coldest flight I;ve ever been on, to 28C. Perfect. Spectacular forest on the way in - can't wait!
View from the taxi. Not too shabby!

Friday, 25 September 2015

Here's looking at Kew

Something slightly different this week, with some scanning at the Royal Botanic Gardens at Kew. I was contacted by a former UCL colleague and alumnus who now works in the GIS UNIT at Kew, where they are interested in capturing the structure of some of their oldest and more unusual trees. In part this may be to preserve them, if only virtually, for the record. But more immediately, the team there are exploring new ways to monitor and manage old trees for their own health and that of the visitors - large falling branches are potential hazard! We went out to Kew and scanned one of the oldest trees in the gardens, an oriental plane, ancestor of the London plane, planted in 1762 (or perhaps earlier) and now rather venerable, hollow, and listing in places. We also scanned the renowned stone pine, planted by Princess Augusta in 1846, and kept in essentially bonsai conditions. It's not little any more as you can see below and the odd shape is a result of the early pruning, but of course leaves it very prone to falling over and breakage.

Bonsai stone pine, with guy wires in the centre. 170 years old, 14m high and rather elegant.
Oriental plane, planted 1762 (perhaps earlier), before the sun came out.

Heavily downsampled point cloud from 3 scan locations, of the stone pine.

As above, but this time with Andy's defoliation algorithm applied, and the remaining wood-only parts coloured red.
Hopefully we'll be able to get back and scan the Plane leaf-off, after some remedial surgery has been carried out. We should then be able to see what the possible effects are on the remaining structure, if any.

A very civilised place to work: you can scan lovely trees, get a very good coffee and a cheeky croissant and even chat to some of the visitors! Fieldwork at its most genteel.

Thursday, 17 September 2015

Nature's Greatest Survivor: the oak tree doc gets aired!

So the Oak tree documentary is finally going to get shown - looks like 1/10/2015. Here's the trailer - and you can see a quick snapshot of one of our scans about 35s in. Exciting!
There's also some of the drone footage shot while we were there, and it all looks beautiful - as it should I guess!

Thursday, 10 September 2015

Some initial results from Kim's work at Wytham

Kim has been working hard on co-registering the 370+ scans we collected from Wytham over the summer. He's managed to pull together the scans and put them into some sort of order, along the LAI2000 and 2200 data and everything else - great job! The co-registration of the scans looks pretty good so far, but is likely to improve as we use the multi-station adjustment on it. But Kim's first fly-through already looks VERY impressive - this includes scans made from the Wytham walkway and so gives a great view up and around the canopy.

As far as I know I don't think anyone has ever scanned as large a chunk of forest as this in such detail so far. The next steps will be pulling out and reconstructing the individual trees, and comparing the resulting models with the inventory data, comparing the structural information with the other ground-based measurements (PAI, LAI, gap fraction), looking at clumping, and then building the canopy radiative transfer model that will allow us to look at fAPAR and satellite signals.

An interesting corollary is the density of targets the individual scan locations 'see' in the plot. Kim has plotted the scan locations with reflectors:
6 ha plot at Wytham Woods, with the scan locations (red) and the reflectors (crosses) marked.
Each reflector is seen from a minimum at least 2 scan locations minimum, and up to 25 maximum, but mostly between 5 and 6 scan locations. This is a useful way to think about how many reflectors we need, the density of scans and so on, as we do this in more locations.

Thursday, 20 August 2015

NERC GREENHOUSE project Summer School

As part of my involvement in the NERC-funded GREENHOUSE consortium, led by Mat Williams at Edinburgh, Andy and I were scanning at Harwood Forest in Northumberland this week. The NERC GREENHOUSE initiative covers 3 different consortia, attempting to provdie robust estimates of GHG fluxes over the UK land surface and coastal and ocean regions, from top-down, bottom-up measurement and modelling schemes.

The Summer School brought together 25 or so NERC-funded early career scientists (Postdocs and PhD students) from across the broad area of GHG research (land, atmospheres, oceans), for a week of demos, tutorials, lectures and practical sessions to showcase @NERCScience across land, ocean and atmosphere. We were showing the students how we can use terrestrial laser scanning to estimate above ground biomass, and forest structure and function, in the Sitka plantations at Harwood, alongside FR colleagues, Mike Perks, James Morison et al. The sun shone for at least one day, we got some scan data (see below) and it's another environment to try out our TLS methods (slightly less challenging logistically than Brazil, and slightly more homogeneous. Slightly).  It was a bit of blast from the past for me, having worked there a fair bit as a NERC Postdoc myself during the old NERC Centre for Terrestrial Carbon Dynamics. It was good to be back in such a lovely landscape; the Cheviot Hills and surroundings really are a stunning and remote, wild landscape. I took my bike with me and went out for a couple of lovely spins through the hills, rabbit-related injuries notwithstanding (honestly: a rabbit ran through my bicycle front wheel, throwing me over the bars and leaving me with fractured clavicle, 2 chipped vertebrae and numerous other painful contusions. Pesky wabbit - RIP).
View from the bike: Coquetdale, early morning.
The students seemed to enjoy the opportunity to engage with the field measurements and asked a lot of interesting questions, even though for some of them we were quite a long way from the sea.
Looking along the stand at Harwood towards the flux tower.
Andy scanning from the road, with the 35-40 year aged stand we are measuring to the right.
Sunshine! And a well-earned lunch break for the summer school gang.
Initial results show we are able to pull out the trunks up to the crowns very well, scanning at a slightly higher resolution than previously (0.02 deg). But the rain on day 2 meant we could only scan at 6 locations, so the small area of plot meant we had less detail from the upper part of the crowns than we have had for other, larger plots.
Oblique view of scanned plot, with the individual trees pulled out and coloured by height up to max of ~30m. 

Looking down on the canopy showing the crowns, again coloured by height. The trees are very similar in height and so the general trend of higher to lower (red to green) from left to right is a function of the underlying topography.
See up-to-date details from @GHGProg on Twitter.

Thursday, 16 July 2015

Exciting times ahead

So I've just heard that we (myself, Simon and Lewis) were awarded the NERC grant we applied for back in January this year "Weighing trees with lasers: reducing uncertainty in tropical forest biomass and allometry". I'm still awaiting the final confirmation that we got the budgets we asked for and so on, but this will provide us with a postdoc for 3 years, and travel money to deploy the TLS across 3 continents, at 4 or 5 contrasting tropical forest sites. The project will enable us to scan several 1000s of trees and estimate above-ground biomass, as well as look at structure and size of trees across contrasting environments and climate regimes in the tropics. We will get to revisit sites in Gabon, and Brazil to see how they've changed in 2-3 years, as well as several new (to me anyway!) sites in Peru and SE Asia. We've also got a range of really exciting collaborations lined up, academic, industry and for our impact plans to bring our research to a range of schools, as well as new visualisations of the 3D data. We'll also be contributing directly to new satellite missions through calibration and validation datasets, most notably the ESA BIOMASS and NASA GEDI missions.

So, some exciting times ahead. And a valuable lesson learned: avoid writing proposals with 2 co-I's with the same surname  - it's very confusing. 

Tuesday, 30 June 2015

Bluest of blue sky

End of June and we had an absolutely stunning day at Wytham. Mike and Zack from Skycap came up with their DJI Inspire and filmed the standing oak for the BBC documentary; we got some lovely footage around the tree which we'll see in the finished piece. Then we flew up from the site where Olli and Teemu from the Finnish Geospatial Institute (FGI) of the Finnish National Land Survey (NLS) have been based the last few days with their very impressive custom-built hexacopter. They have a multispectral imaging camera, with a gimballed cosine head downwelling irradiance sensor (for those who want to know!) which allows them to capture quantitative measurements of reflectance, not just pretty pictures. The first look at the Skycap footage is pretty stunning.
30 C and blue as the sea - the sky is always like this in the English summer - in my mind at least. I was very impressed with how steady the camera gimball kept the view regardless of the drone attitude. You can see the flux tower from around 0:20 s in, towards the top centre of the wood. The aim is to try out height model software, compare with the NLS data, and the Riegl canopy height model data.

Wednesday, 24 June 2015

Wytham Woods, summer 2015

We've been working in Wytham Woods, just outside Oxford, over the past few weeks, scanning and characterising the forest structure in a 6 ha (300 m x 200 m) area. This is the the largest single area we've scanned so far - quite an ambitious task, and is part of our work on the EU-funded METEOC-II project. 

Although the forest is not as dense or tall as the tropical forests we've encountered in Gabon, Brazil and elsewhere, it's still a challenge. Kim has been leading the team (Niall, Ally et al. from NPL, along with Toby and others from Yadvinder's group at Oxford) in setting up the plot, marking out the 176 (!) separate 20 x 20 m grid points, laying out targets and then doing the 352 resulting scans.

Following the lidar scanning, we're collecting LAI and hemiphoto measurements at all the grid points in the plot. The aim is then to be able to use the lidar data to generate a detailed 3D model of the whole site - something that's never been done before on this sort of scale - both leaf on, and leaf off. We'll then be using the model to simulate satellite observations, as well as ground-based observations (hemiphotos, LAI, fAPAR etc) with the aim of being able to provide traceable uncertainty from ground mesrements through to satellite observations. This is essentially impossible to do in 'real life' because you never know what the real LAI or fAPAR of a site like this is. So we're using the 3D model to simulate this process.
A rather lovely view from under the canopy in Wytham, showing the rather beautiful sun-dappled understory.
The woods at Wytham really are a lovely place to work, and very convenient. We're able to visit quickly, and there's a lot of other work going on with various groups, meaning there's good infrastructure. And good pubs and coffee within a 10 minute drive - essential, particularly if you're working there for several weeks at a time.

The Finnish Geodetic Institute UAV coming in to land.
Meanwhile, our colleagues from Finland have been deploying their UAV, with a gimbal-mounted hyperspectral camera on board, to collect spectral data from above the 6 ha site. This will tell us about what's going on in the upper part of the canopy. They have a very nice set-up on the UAV, with a cosine-weighted irradiance sensor (also gimbal-mounted) and some portable calibration targets on the ground. An impressive bit of kit. The video shows it coming in to land.

The FGI UAV. Note the irradiance sensor on the very top with the white housing.
The last of the 352 scans. Kim is happy.
Kim, Ally and Niall (L to R) try and get to grips with the rather primitive menu on the LAI2000. It's not iOS, for sure.

Wednesday, 29 April 2015

Lidar workshop in João Pessoa, Brazil

I've spent the past few days in João Pessao on the very Eastern tip of Brazil - the city where the sun rises first! I can certainly confirm that the sunrises are pretty spectacular - swimming off the beach from my hotel at dawn is pretty special, particularly looking back at the weirdly retro/futurist James Bond villain hotel.
Seeing the sun first in all the Americas!
The Hotel Tropical Tambau, with missile silo, shark tank and concrete nuclear bunker.
I was invited as part of the XVIIth Brazilian Symposium on Remote Sensing (XVII SBSR), a biennial science meeting for Brazilian agencies and specialists in remote sensing. The Brazilian government has invested heavily in remote sensing infrastructure and training for environmental monitoring over the last two decades, particularly in an effort to monitor and control deforestation and degradation, but more widely to support mapping, commercial and conservation activities. I was invited by Dr. Veraldo Liesenberg, a lidar expert from the Department of Forest Engineering (DEF), Santa Catarina State University (UDESC), to come and help lead a workshop in lidar remote sensing for forest mapping and monitoring, in conjunction with Dr. Akira Kato from the School of Horticulture, Chiba University, Japan. The workshop was run over the weekend prior to the main meeting (along side a range of other courses) and was aimed at showcasing the utility of lidar, both TLS and ALS, and in combination with image and other data, for addressing forest science. 
The meeting was held at the very new and shiny "Centro de Congresso de João Pessoa" of João Pessoa city, Brazil. The local authority have built the centre to attract large conferences o the region.
The workshop seemed to go down well and we had around 20 attendees who were mostly students, but also some postdocs, in Brazilian universities, all interested in what lidar could do for them. We presented new results from our work in ALS and TLS, and then ran a series of practicals looking at lidar data and some of the free tools that can be used to handle them, including Cloud Compare, Fusion, WEB-lidar, libLAS and so on. It was definitely an interesting meeting and I met a number of people with interests in our TLS work, and who potentially have plots and data that would be mutually beneficial. 
Everyone wants UAVs. And lidar.
There was also a large commercial expo as part of the conference, showcasing some very interesting new developments in UAV and ground-based tech. UAVs appear to be particularly attractive here due to the relatively lower cost of entry than many other, piloted systems. And the instrumentation and software is obviously advancing rapidly to allow exploitation of these platforms. 

Wednesday, 22 April 2015

New ESA Massive Open Online Course (MOOC)

The new ESA Massive Open Online Course (MOOC) I've been involved with has now gone live. The MOOC is a free, 5 session course of 2-3 hours a wekk, providing an introduction to monitoring of climate with satellites. It's aimed at policy makers, students, and anyone with a general interest in climate and so it's not technical. I was involved as one of the presenters, but don't let that put you off - some of the other bits are excellent. Please circulate to anyone you think might be interested. There's some really interesting bits on the solid earth, fires, the cryosphere, oceans, atmosphere and then other services which depend on EO data.

The course information page is here and you can see some clips from our trip to the Amazon in the trailer below (e.g. at 2:19 - 2:30).

Thursday, 5 March 2015

TLS demo for students and a nice view of the UCL quad

We took the TLS instrument out into the quad here at UCL yesterday, to demonstrate the capabilities of the instrument for collecting 3D data, and what you need to do in order to process those data into something useful. Thanks to Andy and Kim who set up the demo - you can recognise various people in the scan if you know them! They cast long shadows in the UCL quad (literally, rather than metaphorically, although ......).

Monday, 16 February 2015


The Reuters piece has been slightly extended and re-edited with some different material from me, and some additional animations (not from me) resulting in the following:
The piece makes a few different points, mainly about the methods, but also discusses the Brazil work a little more.

Tuesday, 27 January 2015

Short Reuters interview

The short interview I did for Reuters Technology is up. It showcases the recent Australia work in Kim's MEE paper but also the initial flythrough from Caixuana.

ESA BIOMASS mission science meeting

This week at ESA's Earth Observation HQ (ESRIN) in Frascati, scientists with an interest in measuring forest C stocks and biomass more generally, are meeting to discuss the science aims and challenges of the ESA BIOMASS Earth Explorer mission. BIOMASS will make routine observations of P band (70 cm) radar, which is related to the amount of standing C in forests. 
ESA's BIOMASS mission: a slightly dodgy artist's impression (from ESA)
The meeting is to start thrashing out what exactly what we need to do to make the most of BIOMASS observations. In particular, there is a big focus on calibration and validation activities - BIOMASS will rely on calibration of the radar signal against other measures of C stocks, particularly from ground plots, but also airborne lidar. These will all rely in some way or another on allometry, and this is where the TLS comes in. There is a huge need for accurate estimates of C stocks from plot data, hence my invite to come and show what we can do and how this might be a very effective way of augmenting plot data and linking to airborne lidar for cal/val. This is an interesting and varied meeting, as it brings together a wide range of including forest and conservation experts, ecologists, lidar, radar and other EO specialists, modellers and so on. Good to see Ralph Dubayah here, long-tome lidar champion and PI of the NASA GEDI mission to be launched on the ISS around the same time as BIOMASS (2019), and which will make highly-complementary measurements.