Here are some results for MARGINAL reconstructions:

((ta:0.9,tb:0.9):0.5,tc:0.9); with JTT

M3L: mean_accuracy= 0.558 sem= 0.017

LAZ mean_accuracy= 0.537 sem= 0.016

((ta:0.1,tb:0.8):0.3,tc:0.6); with JTT

M3L: mean_accuracy= 0.896 sem= 0.009

LAZ mean_accuracy= 0.895 sem= 0.009

((ta:0.2,tb:0.2):0.1,tc:0.3); with JTT

M3L: mean_accuracy= 0.937 sem= 0.008

LAZ mean_accuracy= 0.938 sem= 0.008

M3L: mean_accuracy= 0.944 sem= 0.008

LAZ mean_accuracy= 0.949 sem= 0.008

((ta:0.8,tb:0.8):0.4,tc:0.7); with JTT

M3L: mean_accuracy= 0.611 sem= 0.016

LAZ mean_accuracy= 0.61 sem= 0.016

M3L: mean_accuracy= 0.614 sem= 0.016

LAZ mean_accuracy= 0.61 sem= 0.016

((ta:0.1,tb:0.1):0.1,tc:0.1);

M3L: mean_accuracy= 0.986 sem= 0.003

LAZ mean_accuracy= 0.986 sem= 0.003

M3L: mean_accuracy= 0.98 sem= 0.004

LAZ mean_accuracy= 0.976 sem= 0.004

M3L: mean_accuracy= 0.987 sem= 0.003

LAZ mean_accuracy= 0.986 sem= 0.003

((ta:0.05,tb:0.15):0.1,tc:0.15); with JTT

M3L: mean_accuracy= 0.984 sem= 0.004

LAZ mean_accuracy= 0.983 sem= 0.004

M3L: mean_accuracy= 0.977 sem= 0.005

LAZ mean_accuracy= 0.977 sem= 0.005

M3L: mean_accuracy= 0.974 sem= 0.005

LAZ mean_accuracy= 0.973 sem= 0.006

This data suggests to me that my M3L implementation of marginal ASR is more accurate than PAML’s implementation.

More evidence, now with a gamma model.

((ta:0.7,tb:0.7):0.2,tc:0.7);

M3L: mean_accuracy= 0.706 sem= 0.005
LAZ mean_accuracy= 0.7071 sem= 0.005

M3L: mean_accuracy= 0.7041 sem= 0.005
LAZ mean_accuracy= 0.7067 sem= 0.004

M3L: mean_accuracy= 0.7521 sem= 0.004
LAZ mean_accuracy= 0.7545 sem= 0.004

I just returned from a quick overnight hike to Benson Lake in the Mount Washington wilderness (see hike #46).  As of July 15 2011, the snowline is at about ~5000 feet, but the melt is very uneven.  Many south-facing slopes are snowfree, especially above tree line.  The bald southern face of Scott Mountain (6100′) is entirely melted.  Black Crater (7251′) and Belknap Crater (6872′) appear to have snowfree hiking routes above treeline, but I suspect their approach hikes are snowbound.  North facing slopes and deeply shaded forest areas have four-foot drifts all the way down to Highway 242 (at about 5000′).

The Benson Lake trailhead (-5000′) remains unreachable by car.  Instead, I parked about 1/2 mile down Scott Road.  I found the trail to Benson Lake (5200′) about 80% snowbound, but I was able to follow good boot prints from early-season day-hikers.  Benson lake basin was mostly snow-free; I found a dry campspot with a view and built a fire.  Silent lightning flashed for hours in the southern sky , and later, rain clouds consumed me, Benson Lake, and the entire wilderness.

Check-out more photos on Flickr.

Summary: Here is a novel route to access Iva Bell Hot Springs in the John Muir Wilderness of the Sierra National Forest.  Most people reach Iva Bell from the north (via Red’s Meadow), or from the east (via the PCT).  However, in heavy snow years—such as this year, 2011—the north and east trailheads may be unaccessible early in the season. The route I describe here provides early-season access to the hot springs from the west, starting at the McCreary trailhead in the Ansel Adams Wilderness.  CAUTION: This route uses trails that are poorly maintained.  Navigation skills are required, especially to traverse the section from Heitz Meadow, around Pincushion Peak, and down the Silver Creek drainage.

Trip Statistics: Approximately 25 miles one way.  About 6000′ total gain one way.  McCreary trailhead is located at 6774′. Cassidy Bridge is the lowest point at 4400′.  The saddle north of Pincushion Peak is the highest point at 8622′.  Iva Bell hot springs are located at about 6450′.

Conditions in June 2011: Small snow patches lingered on the Minarets Road, but we were able to drive a sedan within one mile of the McCreary trailhead.  The trail to Cassidy Bride was completely snow-free and easy to follow.  There were about a dozen poison oak patches on the trail at the bottom of the San Joaquin river canyon; we climbed around these patches, but pants and a long-sleeve shirt are recommended for this section.  The switchbacks east of Cassidy Bridge were in good condition, but the trail beyond to Heitz Meadow was sometimes difficult to find.  Beyond Heitz Meadow, the trail disappeared under fallen trees and forest litter. Snow remained above ~7500 feet, and we hiked for several miles on solid snow. String Meadows was completely snowbound. Overall, we trekked essentially off-trail from Heitz Meadow, over the saddle north of Pincushion Peak, and into the Silver Creek drainage. The terrain was amenable to off-trail travel, but the forest is mostly viewless and good navigation skills are required—a GPS device could be useful here!  There was an excellent log crossing over Silver Creek at the same point where the trail crosses the creek.  Without the log, Silver Creek would have been too full to safely ford.  The switchbacks down to Fox Meadow were easy to follow, but they were overgrown with manzanita and huckleberry oak; this is another section where long pants are recommended to protect your legs from sharp branches.  At Fox Meadow, we easily found the well-maintained trail to Iva Bell hot springs.  We may have been the first people to soak in the springs this season.  The meadow pools were covered in algae, but we easily scooped away the mess and enjoyed the soaking.  The upper pools, unfortunately, were filled with hundreds of tiny leeches; I think these pools are not soakable until the leech problem is solved.  We camped at Iva Bell for two nights.  We saw another small group that entered over Mammoth Pass.  The shuttle to Red’s Meadow was not yet running during our trip, so I’m not surprised that the area was basically empty.  We returned using our same route.

A note about navigation: We carried topographic maps, a compass, and an iPhone with GPS.  We relied on the map and compass to navigate the cross-country ascent north of Heitz Meadow, but we also used the iPhone a few times to validate our decisions.  The iPhone GPS, surprisingly, worked very well in the wilderness.

Map: view the route in Google Maps:

Video Postcards from the trip: 

Jefferson Park, September 2009

Today I saw a captivating slideshow by William Sullivan, based on his book “Oregon Favorites”.  (Sullivan is best known for his “100 Hikes” series of trail guides, which are really excellent.)  Sullivan has a busy year-round lecture schedule.  Today’s slideshow featured one favorite place in Oregon for each month of the year, with an emphasis on places that are rarely visited.  Here is the list. . .

March – The Mulino Flour Mill is not a wilderness hike, but the water-powered flour mill still functions and is interesting the visit.

April – The Badlands loop, in the Oregon Badlands Wilderness near Bend, OR

May –  Cape Horn in the Columbia River Gorge

June – The Scout Camp Trail, near Crooked River Ranch, OR

July – Visit the Wallowa Mountains, near Joseph, OR.  Avoid the crowded trails, and visit McCully Basin.

August – Pinnacle Ridge on the NE side of Mount Hood.

September – Jefferson Park

October – Sisters Rocks, on the Oregon coast between Port Orford and Gold Beach

November – Eight Dollar Mountain along the Illinois River.

December – rent a fire lookout tower.  Sullivan recommends Warner Mountain.

January – visit Fort Hoskins Historic Park

February – ski around Broken Top, using the pass/notch near Iceberg Lake

This function is supposed to return a random value from the Gaussian distribution with a mean = 0.0 and a user-specified standard deviation.  However, consider the C code (below), in which the user-specified standard deviation value seems to have no effect on the random value.

Code:

#include
#include "gsl/gsl_rng.h"
#include
int main( int argc, const char* argv[] )
{
printf( "\nHello World\n\n" );
gsl_rng *rng;
int random_seed = (int)time(NULL);
rng = gsl_rng_alloc(gsl_rng_mt19937);
gsl_rng_set(rng, random_seed);

double rand_gauss;

int reps = 5;
int i = 0;
for (i = 0; i < reps; i++)
{
rand_gauss = gsl_ran_gaussian(rng, 0.001);
printf("sigma = 0.001, random value = %f\n", rand_gauss);
rand_gauss = gsl_ran_gaussian(rng, 0.01);
printf("sigma = 0.01, random value = %f\n", rand_gauss);
rand_gauss = gsl_ran_gaussian(rng, 0.1);
printf("sigma = 0.1, random value = %f\n", rand_gauss);
rand_gauss = gsl_ran_gaussian(rng, 1.0);
printf("sigma = 1.0, random value = %f\n", rand_gauss);
fflush(NULL);
}
}

Result:

Hello World

sigma = 0.001, random value = 6304.000000
sigma = 0.01, random value = 3104.000000
sigma = 0.1, random value = 1952.000000
sigma = 1.0, random value = 1536.000000
sigma = 0.001, random value = 3712.000000
sigma = 0.01, random value = 1312.000000
sigma = 0.1, random value = 3424.000000
sigma = 1.0, random value = 6752.000000
sigma = 0.001, random value = 6016.000000
sigma = 0.01, random value = 8000.000000
sigma = 0.1, random value = 6752.000000
sigma = 1.0, random value = 7200.000000
sigma = 0.001, random value = 5408.000000
sigma = 0.01, random value = 7136.000000
sigma = 0.1, random value = 1056.000000
sigma = 1.0, random value = 4160.000000
sigma = 0.001, random value = 7744.000000
sigma = 0.01, random value = 5376.000000
sigma = 0.1, random value = 3072.000000
sigma = 1.0, random value = 4896.000000

WTF libGSL?

If you see this error. . .

RuntimeError: maximum recursion depth exceeded

. . . don’t freak out.  Instead, try increasing the recursion limit.  First, find the current limit:

>>> import sys

>>> sys.getrecursionlimit()

1000

. . . and then increase the limit:

>>> sys.setrecursionlimit(10000)

In my case, I changed the limit from 1000 to 10000, but obviously the appropriate numbers might be different for your problem.

Contemporary anti-foundationalism is rampant—for example, consider recent skepticism about global warming, child vaccination, and evolution—and it’s easy to think that the counter-enlightenment is a relatively new phenomenon.  However, Christian Thorne points out in the introduction of his book that the counter-enlightenment movement has existed as long as the enlightenment movement—the counter-enlightenment has been the dark shadow of the enlightenment.  The enlightenment has not been a constant march towards knowledge, but rather a fitful cycle of awakening and forgetting.  Consider this quote from Thorne’s introduction:

The period’s conventional history designations, for England as for Western Europe more generally, seem to tell a single story: the Renaissance, the Revival of Learning, the Great Renewal, the Scientific Revolution, the Enlightenment, the Age of Reason. These names all conceive of Europe’s early modernity as a philosophical awakening. . . and yet there is something curious about these terms all the same, since on a second look, they don’t so much tell a single story as they tell the same story over and over again.  These designations, after all, may all come under the umbrella of early modernity, but they don’t really refer to the same period at all.  Reduced to their historical caricatures, they name distinct, if marginally overlapping, blocks of time: the fifteenth- and sixteenth-century Renaissance (extended in England to the civil wars), the squarely seventeenth-century Scientific Revolution, the eighteenth-century Enlightenment.  And what’s strange is this: these aren’t merely terms from intellectual history, christening each century after its distinctive doctrine of world view.  Each term suggests that its period was *newly* intellectual, *newly* philosophical, that knowledge began *then and there*, or at least began anew—as though knowledge needed first to be reborn, then revised, then renewed, then revolutionized, and only then, slowly, would enlightenment dawn.  These terms, arranged in series, suggest less the triumphant emergence of reason than an embarrassing intellectual fitfulness, less a scholarly awakening than a philosophical narcolepsy, so that we must imagine Europe dozing off, repeatedly, only to freshly awake after each new nodding.