The benefit of the drought

By some estimates, California is experiencing its most severe drought in more than a millennium. I get asked a lot about the drought-tolerant plant installation I created last year, and so I’ve done a little write-up in the hopes that it might inspire others in their water-conservation efforts.

I took advantage of the Los Angeles Department of Water and Power’s turf removal rebate, which offered $3.75 per square foot of turf replaced with water-friendly landscaping. Although the state funding for this program has mostly, er, dried up, a limited number of agencies including LADWP are offering rebates at the time of writing.

Once upon a time we had a lawn, but with the water shortage the cost of maintaining it was starting to weigh heavily – on the wallet, on the environment, and on the conscience. Following the bone-dry SoCal winter season 2014-15, it didn’t look too great. I decided it was time.

I applied for the rebate around March 2015 and stopped irrigating the lawn. It took about 3 months to get approval; here’s what it looked like after 3 months of no water:


Horrible. This was my starting point.

Although there are other ways to kill off the remaining grass, I decided to rip it all out. I chose this option because the soil had become rock-hard, and I felt it could use some loosening and aerating before attempting to plant anything. So I started to dig…

I quickly found that our landscapers had done something a bit dodgy with the low-voltage cables feeding the outdoor lighting. Instead of burying them to the proper 18-inch depth, they had merely draped them on top of the soil and hidden them with the turf. My first task, then, was to bury these wires.


As I continued loosening soil, I started to encounter the buried irrigation pipes. One of the requirements of the rebate program is that pop-up sprinklers used to water the lawn be removed, and optionally replaced with low-water outlets – microsprays or a drip system. I had planned to put in a drip system and wanted to make use of the existing irrigation pipes, so when I accidentally broke a couple, I needed to fix them. Which I did. Then I discovered the pipes I had fixed were part of an older irrigation system no longer connected. Nevermind, it was good practice. It also inspired me to make a map of the whole area showing all PVC pipes, electrical cables, drain pipes, water main and so on.

By this point, I was beginning to wish I hadn’t started:


But, eventually, everything was nice and loose, and mapped-out, and buried again.


I now had 1225 square feet of bare earth surrounded by wiggly edging, as my blank canvas. For me the hard part is knowing what to put there. To this end I had spent many months trying to gather ideas from the web and from local gardens, in an effort to decide what I liked and might be capable of emulating. I opted for a multi-texture approach, and planned to cover the area with a mix of wood-chip mulch and crushed rock in two or three colours.

Mulch. What an excellent word. Decorative mulch is available at many outlets, but plants really require a 2-3 inch layer, and the cost soon racks up if you’re trying to cover any significant area. But did you know that it’s often possible to obtain free mulch? It won’t look quite as pretty as the kind in the shops, but will perform the same functions – reducing evaporation, and decomposing to feed nutrients into the soil. The way to get it is through a tree-trimming company. They routinely generate tons of mulch on a daily basis, and some of them will drop it off at your house for free. Here’s the company I used:

The only drawback – they couldn’t deliver anything less than a truck-load. So that’s what I got:


Of course I had to get it off the road. 27 hours later, and 44 trips with a 60-gallon wheelie bin, I had moved it round the back of the house. At roughly 100 pounds per bin-load, that’s 2 tons of mulch, and needless to say, I was in pain for a while.

I also ordered about 3 tons of crushed rock and crushed red brick from a local supplier – Central Valley Builders Supply (, which got delivered onto my driveway. Oh good, I thought, more shovelling.


I capped off all the sprinkler risers. A tip here: to prevent leaks caused by mains water pressure, use a half-dozen turns of teflon tape on both threads of the riser, hand-tighten, wait a day, then hand-tighten again before burying. If you need to be able to remember where they are, for future use, mark their locations like I’ve done here with blue flags:

IMG_1559 (2)

I also had some nice flagstones just lying around, not really being featured, so I experimented with shaping them into a stepping-stone pathway. This immediately started giving some structure to the area.

Part of the plan was to create a border using the crushed red brick, which I began delineate using plastic edging:


Now, the rebate program requires that a weed barrier (landscape fabric) be installed over the entire area. I can think of 3 reasons why this might be so – it traps more moisture by reducing evaporation, it eliminates the problem of weeds interfering with any new plants which are trying to establish themselves, and lastly a rather human reason – it makes homeowners a bit less inclined to re-install grass at a later date. The weed barrier can be seen going in here under the red brick:


Boulders make a nice enhancement. The one on the left was repurposed from another part of the garden; the one on the right I actually dug up when loosening the soil! This soil has a lot of clay, which becomes compacted over time, some of it eventually forming these crumbly bouldery chunks. I found a couple of other nice rocks on countryside walks.

The landscape fabric must be fastened with staples. The rather flimsy 4″ ones from the local garden centre I found to be a bit too insubstantial for the job, so I ordered some chunkier 6″ ones from a pet supplies seller on eBay. I suppose they’re for fastening dog fences but they work perfectly here, and are very cost effective when ordered in bulk – about 5c per staple.

With the border complete I used ropes to experiment with shapes for the other rock areas:

IMG_1624 (2).JPG

After finalising a shape, I would mark it onto the soil and use the line as a guide for laying the edging. I found I got quite good at doing the curves after a while, although to certain degree the edging tends to fall into something akin to natural spline, due to constraints imposed by the soil pushing on the buried section.


After giving this area the weed barrier treatment, some tan coloured stones went down:


I received plenty of assistance at every stage from furry friends.


The juxtaposition of mulch and stones provided the textural contrast I was aiming for:


The mulch must go on top of the weed barrier. How then, I wondered, does it decompose and feed the soil? My friend Paul Van Middlesworth explains – the feeding occurs at the microscopic level, so once the mulch is sufficiently decomposed, when it rains the nutrients simply filter down through the weed barrier.

I created another region for the white stones:



I continued laying weed barrier and covering with mulch, until done.




Fully mulched. Time for some plants! To qualify for the rebate, a minimum of 40% of the area must be covered by water-friendly plants once they’ve reached maturity. I ended up with about 50 plants in different varieties, most of them quite small (1-gallon pot or thereabouts) as I wanted a fairly low growth height over the whole area, and also to keep the budget down. Selecting/purchasing/planting took the better part of a month, and I ended up with the following plants, all of them drought-tolerant to varying degrees:

  • Purple fountain grass
  • Mexican feather grass
  • Lantana (yellow, orange/red, purple)
  • Carmel Creeper (a california native)
  • Creeping Mahonia (a california native)
  • Greenmound Juniper
  • Rosemary (transplanted from the back garden)
  • Blue chalk sticks (succulent)
  • Cabbage-head agave
  • Pink ice plant
  • Agave americana marginata
  • Caribbean agave
  • Echeveria

These choices were influenced to an extent by my having done some research into what grows well in my climate zone, but mainly involved just picking out what looked nice in the low-water sections of the local nurseries.

It’s surprising how many plants and compost bags you can fit in the back of a smart car:


I’ve since been inspired to do my own composting, so I shouldn’t need to buy compost next time around.

Planting entails the following. First, brush the mulch (or rocks) away from where the plant is to go. Then, cut an ‘X’ shape in the weed barrier, large enough so it can be folded back to accommodate the hole you’re about to dig. Dig a hole as deep as the pot, but twice the diameter. Remove the plant with its root-ball and soil, and place into the hole. Now, using some of the soil you dug out of the hole, mix it 50/50 with compost, and fill the gap with the mixture up to the level of the ground.


Fold the weed barrier back into place and cut away any part of it which collides with the base of the plant. Replace the mulch. Water thoroughly, and then water ever day for a week or two until the plant has established roots, after which you can start to cut back the watering, tapering off to the target amount for the plant.

Repeat, per plant.




The requirements satisfied, I submitted the project completion form, but I was still forced to do a great deal of hand-watering while plants established themselves. What I really needed was a drip system, so I turned my research efforts to this next.

Our mains water pressure is about 120 psi, which is much too high for drip components. So one of the elements of a typical drip system is a pressure regulator that reduces pressure to the level required by the down-stream components, typically about 25 psi. However, the pressure regulator for a drip system is generally made of plastic, and you may find that it ruptures after a while when subjected to mains pressure. On the advice of my friend Paul again, I installed a proper brass pressure regulator – not quite as heavy-duty as the one that supplies your house, but still robust enough that it won’t fail under mains pressure, and with an output range adjustable from 10 to 70 psi.

I installed the regulator upstream of all the irrigation valves – not just the one controlling the drip system, but also the ones controlling the existing sprinklers. Mains pressure is usually too high for ordinary pop-up garden sprinklers, which explains in part why they crack from time to time and need replacing in the absence of regulated pressure.

I had this region of space to fit my installation in, between the copper pipe and the first valve:


Based on Paul’s suggestion, I used shut-off valves to set up a 2-way flow path, one for full mains pressure and one for regulated pressure, so it’s easily possible to switch between them for any reason. Here’s the layout I managed to squeeze into the available space:


With reduced pressure now available, I began work on the drip system itself. I found that there are various kinds of manifold which convert a half-inch sprinkler riser so that you can feed quarter-inch drip tubes with it. I used an 8-way kind, made by Orbit, called Apollo-8:

IMG_2005 (2).JPG

Each of the outlets on the manifold can be individually adjusted with a screw, and/or capped off completely. 5 of these manifolds provided more than sufficient capacity.

I ran quarter-inch drip line to the base of each plant, terminated mostly with 1 gph (gallons per hour) nozzles. I also had 0.5 gph and 2 gph nozzles on hand, which I occasionally used depending on the plant type.


Lines were branched using T-junctions and secured with stakes.


I arranged two nozzles to point to the base of every plant, watering it from opposite sides.


When the lines are then covered with the mulch, the entire drip system is almost invisible. Remarkably, I even came across some white and tan-coloured drip line, which I used for camouflage against the stones. (Ok, maybe that’s overkill – the tubing is buried anyway.) I drilled through the plastic edging to feed water into the stone areas:




I completed the drip installation in 2 days, during which time I installed over a hundred nozzles plus the associated junctions and stakes. The small angular nature of the plastic components, combined with the extreme fingertip force required to securely join them to the tubing, caused me to work my fingertips painfully raw. As a result of this they became leathery and hard after a couple of weeks. Then all the skin peeled off a couple of weeks later. Then the next layer peeled off a couple of weeks after that. I would recommend a somewhat more leisurely pace for drip installation.

It was all worth it, though. I like to think we have a nice-looking front garden where there used to be the dried-up remnants of a lawn.




We eventually received a substantial cheque from the power company that more than reimbursed for the cost of materials. And, now that the plants are well-established, the volume of water required to keep them happy is a tiny fraction of what was once used on the same area.





K’NEX roller coaster

I recently found some old video footage of a K’NEX roller coaster I designed and built around the year 1999. It predated YouTube by a considerable margin and never found its way onto the web, until now. Here it is (please forgive the last-millennium image quality):


It was constructed from about 10,000 individual pieces, and although it was unremarkable compared with the creations of K’NEX superfans or with the more modern K’NEX roller coaster construction sets available, the aspect which made this design an interesting one is that the building set it used was quite basic: the classic K’NEX Roller Coaster 63030 Builder’s Challenge, which featured nothing more than the standard general-purpose rods and connectors augmented with a couple of simple extra parts for making the rails and the chain.

The original 63030 set featured 2,400 pieces with instructions for building two different roller coaster configurations – ‘spiral version’, and ‘loop version’. I recently tried building both of these on the kitchen table (much to the annoyance of my wife):

The box boasts “8 feet long! 3 feet high! 26 feet of track!”. All very impressive, and made for an exciting toy. However, knowing a few things about roller coaster design and the mathematics behind the track curves, the purist in me found himself somewhat dissatisfied. The loop was circular in profile. There was a discontinuity at the base of the lift hill. The curvature did not change smoothly along the track. It was only capable of handling a single car. Above all, it wasn’t nearly big enough!

This led me to investigate the possibilities of pooling several sets to construct something larger which adhered a bit more closely to real-world roller coaster principles. I don’t recall how long I spent tinkering with it, but I would guess it took me a couple of months including all the fine-tuning needed to make it run reliably.

These are the final stats:

  • Height: 6’11”
  • Ride duration: 47 seconds
  • Track length: 79 feet
  • Inversions: 2 (loop, corskcrew)
  • Cars: 3
  • Cars per hour: 180
  • 1st drop: 34″
  • 2nd drop: 21″
  • 3rd drop: 34″
  • Loop height: 18″

Cars first wait in a pickup area until caught by a hook on the chain lift:


They are then pulled by the motor-driven chain through a short tunnel (whose true purpose is to hide excess track tubing!) and up a lift hill, taking 20 seconds to reach the summit:


A motor provides the power to lift cars to the top. The motor was available as an accessory to the original set and appears to work fine with the much longer chain:


A 180 degree turn off the top of the lift hill leads to the terrifying first drop and into the first inversion, a loop:



As in most real roller coasters, both the drop profile and the loop exhibit curvature which changes gradually along the length of the track, increasing steadily to a maximum at the loop apex:


Another turnaround leads to a second, smaller drop:


Riders are then given a brief instant to prepare for another element inspired by real roller coasters – a banked drop into a corkscrew:



The pace then lets up considerably with a slow turn into a gentle 720 degree helix:


This opens out into a section with a long back-straight, which might make a good place for the loading platform of a real-world roller coaster:


Especially challenging was the design of the car. I found it insufficient to copy the design straight out of the K’NEX instruction book, because it lost energy too fast, and could not twist freely enough about its long axis to negotiate the corkscrew element. You can see the high degree of twisting required:


I experimented with several designs before settling on one that worked better for my purposes. The result was a longer car, stripped down to a mere chassis, with more rigidly constructed bogies joined by a looser coupling. Crucially, each bogie features an off-axis steering point which permits a continuous axle to join the left and right wheels. This allows the axle itself to rotate in its bushings, lowering rolling resistance compared to the original design which forces the wheels to rotate on stationary axles.




K’NEX is a mathematically elegant building set. Its basic part are colour-coded rods in a variety of lengths, and a range of colour-coded connectors which snap one or more rods to each other. The ingeniuity lies partly in the lengths of the rods – each rod is √2 times the length of the next shorter one (when lengths are measured from centre to centre of the adjoining connectors), which means that diagonal cross-members can be added to a square lattice.

You might then think that all structures would necessarily turn out looking very rectangular, but there’s another crucial property that allows a curving roller coaster to be built – there’s a certain amount of flexibity in each connector, which means rods don’t have to join at exact multiples of 45 degrees. So a straight track can be made to bend gradually. Furthermore, combining rod lengths in differing ways allows for a range of curvatures. For example, here is how I transition from a large to a small radius of curvature in the loop:


By offsetting the rods supporting one rail relative to those supporting the other, a twist can be put into the track, which is how the corkscrew element is achieved:


I with to thank Jon Steele for invaluable advice and assistance regarding video file conversion, and Shannon Studstill for the entertainment provided by her adorably savage puppy.

And thanks for reading!



Dissecting a UFO video

This post began life as an email, which I felt compelled to write in response to a ‘UFO’ video I was shown, and which a friend then persuaded me to rewrite in the form of a blog post.

Here is the video:

At the time of writing, it has had over one million views in its first month or so.

While I don’t like to spoil the fun – we all enjoy being a bit ‘spooked’ now and again – this video is a great example of the sort of thing most of us could do a much better job of debunking, by improving our critical thinking skills and by doing a bit of legwork.

A first word of advice might be to try not to let the spooky music impair your judgement! Plus don’t be misled by the fact that most of the news presenters appear bamboozled. With the exception of specialist reporters such as science and space correspondents, I find that news presenters quite often lack a good grasp of scientific subjects and of the kind of phenomena shown in the video.

Let’s proceed through the video segments and see where a bit of investigation leads us. I’ll use the elapsed time in the video to mark the start of each segment.

00:11 – China, Jul 7th 2010

An ABC News report tells of an incident in which Xiaoshan airport was temporarily closed when a flight crew saw lights which did not have a corresponding signal on radar. This in itself is perhaps little more than an embarrassment to the airport’s flight traffic control. However, the report proceeds to show us some stunning images, for example this one:

Image from ABC News report on Xiaoshan UFO

The first thing which struck me was that the pictures didn’t appear to match the description. Why weren’t we told something like “residents captured stunning close-up photos of airborne craft flying near Xiaoshan airport?” The pictures also appeared to me to have been taken in very different locations from one another, which at least hints that something might be awry. In situations like this one, TinEye reverse image search proves very useful, and it found the following when applied to the image above:

This is a photo of a helicopter taken in France, 3 years before the alleged sighting. The helicopter is shining a searchlight at the ground, and appears stretched because of the long exposure. It turns out the other images from the news report are also long-exposure photos of helicopters, all of which predate the alleged sighting, with the exception of one which is a similar photo of a plane. More detail is given here:

It is unfortunate, but not atypical for news media, that ABC News decided to jump on the sensationalist bandwagon rather than do a little background checking on the photographs before reporting.

00:45 – China, Jul 15th 2010

We see a Fox News report describing a second UFO sighting in China, and containing video of this beautiful object:

Fox News China UFO

Once again, the image seems to not even vaguely match the description in the report. And, as before, a little digging around reveals that this is actually a video of an earlier incident in a different country – in this case June 30th 2010 in Kazakhstan:

You’ll see that this exactly matches the video in the Fox News segment, but was uploaded to YouTube a couple of weeks before the China sighting occurred. Fox failed to notice that they were showing the wrong video.

Still, we need to explain the Kazakhstan video. What is that luminous trail? Luckily, this effect is a fairly routine sight for anybody living sufficiently close to a satellite launching facility. I live about 100 miles from Vandenberg Air Force Base and have seen several twilight launches which looked just like the object in the video.

In this case, we can research exactly what we’re seeing. On June 30th 2010, an unmanned Soyuz cargo transport called Progress M-06M was launched from Baikonur Cosmodrome in Kazakhstan, to send supplies to the International Space Station:

Here is a video of the launch of Progress M-06M:

It’s easy to see that this is the same object viewed from a different location. The reason the rocket’s exhaust plume takes on that characteristic shape is that the upper atmosphere is very thin and provides almost no drag to the expanding gases, allowing them to expand to huge size. Also, at higher elevations the trail is lit by sunlight while observers on the ground are still in the darkness of night, and this is how the trail appears to glow.

01:25 – Norway Spiral, Dec 9th 2009

We are shown images of a bizarre display. Here’s a particularly nice photo I found on the web showing the same incident, which appears at first sight to have been Photoshopped:

Spiral over Norway

Far from being a mystery, this event was well-known to those who follow science news and astronomy. The cause was quite quickly discovered to be a failed Russian Bulava missile; the third stage of the rocket malfunctioned and began to spin, spewing its exhaust plume like a Catherine wheel:

02:30 – Australia Spiral, Jun 5th 2010

A sighting similar to, though less spectacular than, the Norway one was witnessed over Australia:

Spiral over Australia

You should by now already be equipped to guess at the nature of this object: it’s due to a tumbling rocket stage. The rocket causing it is thought to have been the maiden flight of SpaceX’s Falcon-9:

02:47 – Moscow, Oct 8th 2009

We are shown video of this delightful thing:

Moscow hole punch cloud

Though spectacular, it’s just an uncommon type of cloud formation – you can try an image search ‘fallstreak holes’ or ‘hole punch clouds’.

Perhaps revealingly, the immediately preceding footage of a news presenter describing it has been edited so that it sounds a lot like she says “and scientists / left baffled by…”, where ‘/’ represents the moment of the edit, as can be clearly seen by watching the news ticker jump to a completely different sentence.

03:15 – Jerusalem, Jan 28th 2011

We see parts of 4 videos which allegedly show, from 4 different view points, an object hovering above the Dome of the Rock in Jerusalem’s old city, which then appears to swoop down to a spot just above the dome. It remains there for a while, after which there are some flashes and the object shoots skyward, leaving a pattern of dancing red lights in the sky.

Given the obvious religious significance of the site, we should be on high alert that a hoax might be in effect here. And hoax is precisely what this is. Luckily the debunking has already been done for me, showing that the videos are the result of special effects trickery. (Not particularly good trickery, even.)

The first video has been shown to have been played with digitally. Software has been used to add in a camera-shake effect, introducing gaps at the edges which were filled by mirroring the image contents:

Here’s an analysis of the second video in which camera shake has been removed, revealing severe registration errors between the object and the background:

The third video used this stock photo in place of the real scene, as you can verify yourself:

The fourth video has been debunked by audio analysis:

04:13 New York, Oct 13th 2010

No especially revealing footage is shown, but people in the Chelsea neighborhood of New York spent the afternoon gazing up at some brightly-colored distant objects which appeared to hang in the sky. Here is a representative photo of the incident which I found on the web:

New York Oct 2010

There is no mystery to this one, once it is known that earlier the same day several large bunches of bright yellow balloons were released from another part of the city, in celebration of the Centennial of Madrid’s Gran Via:

balloons released in New York

04:50 – 06:32

The video then launches into a frenzied montage with few specifics and many repeats. While many of the shots are easily identified, I’ll perhaps save this for a future post. However one rather lovely image worth mentioning is this one:

06:08 – Hubble Space Telescope photo, Jan 29th 2010

Hubble Space Telescope captures asteroid collision

This curious display is actually the debris which resulted from a rarely-seen collision between two asteroids:

Following the montage, there are a few more clips worthy of debunking.

06:32 and 06:38 – California, Jan 6th 2011

A set of three lights is shown. In the first clip they appear to hang in the sky; in the second they appear to be falling. A bit of checking reveals that skydivers from the ‘Golden Knights’ US Army Parachute team made a night-time jump using flares over Lake Elsinore, California:

It doesn’t take much imagination to see that the first clip is of the parachutists with their chutes already deployed, and in the second they are free-falling and then deploy their chutes, slowing their fall rate, and begin steering in various directions.

07:16 – Russia, Feb 26th 2010

A mysterious plume hangs in the sky:

Russia UFO Jan 2011

Further searches for this sighting turn up some more UFO videos, such as this one:

But this sort of expanding-plume effect is by now probably familiar to you: it’s a rocket launch, in this case the launch of a Russian Glonass-K1 satellite which took place on that date:

07:22 – Russia, March 7th 2011

We see video of another nebulous formation – here’s an image from the web:

Progress M-03M launch

But the video and claimed date do not match. This is yet another case where a video from a completely different and much earlier event has been trawled up. In this case it is the launch of another rocket, the Progress M-03M supply ship on October 15th 2009, on its way to the International Space Station. (The country at least is correct in this instance.)

Here is some video of the launch:

The rocket appears to be travelling downwards in the video, but this is just an optical illusion: it is travelling away from us and heading over the horizon. It appears to travel down towards the horizon, in just the same way that an aircraft does when travelling away from the observer. And if you are curious about the adjacent objects which appear to flare up at 02:16 in this last video and then separate away, these are the booster rockets being jettisoned. You can see them twinkling later on around 03:30 in this last video, as they tumble in free-fall.

08:15 – Caller to the Art Bell radio show

There isn’t a great deal to add here, except to say that I don’t think the caller will be winning any Oscars. And it doesn’t take much background checking on the radio program’s host to discover that the forum in which this conversation aired is less than credible:

  •  In 1998, Bell was named as recipient of the less-than-prestigious Snuffed Candle Award. The CSICOP Council for Media Integrity cited Bell “for encouraging credulity, presenting pseudoscience as genuine, and contributing to the public’s lack of understanding of the methods of scientific inquiry.”

So, to sum up, am I actually interested in ‘flying saucers’? Yes, I find them fascinating – but as a purely social phenomenon and a tremendously prolific meme (in the original, Dawkinsian sense of the word). There has never been a single piece of scientifically substantiated evidence in support of the notion that we are being or have ever been visited by aliens. But there is plenty to be excited about if you really are interested in the possibility of extraterrestrial life, and a good place to start might be to read up on the following topics:

The SETI program:

The real scientific work behind missions to other bodies in our solar system, for example an upcoming Mars mission:

The hunt for exoplanets, where you can even participate and help find them from your computer: