Meta tags:

Headings (most frequently used words):

january, 2010, the, climate, sunday, sciencapades, 2011, friday, saturday, march, 2012, tuesday, june, warming, heat, island, effect, and, sextant, blog, archive, 25, 11, 14, monday, 13, february, 18, 16, 10, 09, 08, wednesday, 06, 05, 03, 02, 01, about, me, explosion, of, cheney, hate, following, heart, transplant, totalitarianism, all, rage, frustration, trig, letter, where, wi, democrats, unions, don, like, democracy, briffa, finnish, tree, ring, reanalysis, lost, found, population, density, global, seeking, here, problem, science, iv, iii, ii,

Text of the page (most frequently used words):
the (371), and (122), that (91), giss (79), data (76), dset0 (75), dset1 (51), for (51), are (43), with (41), year (40), from (36), this (36), not (33), sta (32), they (31), stations (29), all (27), what (26), urban (23), about (21), temp (21), temperature (20), have (20), station (19), location (19), alt (19), been (18), but (18), over (18), would (18), rod (17), posted (17), then (17), was (17), intercept (17), effect (16), locations (16), other (16), length (16), more (16), one (15), there (15), anomalies (15), means (15), has (15), info (14), how (14), can (14), out (14), hansen (14), into (13), years (13), difference (13), rural (13), effects (13), here (13), january (13), used (13), you (12), climate (12), were (12), plot (12), results (12), did (12), random (12), loc (12), don (11), adjustments (11), same (11), analysis (11), much (11), between (11), temperatures (11), than (11), water (11), vapor (11), even (11), population (11), get (10), these (10), some (10), simple (10), set (10), like (10), mean (10), might (10), 2010 (10), global (10), model (10), increase (10), which (10), density (10), why (9), two (9), given (9), anom (9), names (9), method (9), make (9), when (9), time (9), satellite (9), city (9), heat (8), surface (8), made (8), those (8), using (8), june (8), posts (7), look (7), doing (7), raw (7), island (7), warming (7), black (7), area (7), each (7), next (7), degrees (7), previous (7), right (7), looking (7), found (7), statistical (7), lxs (7), ranef (7), first (6), through (6), since (6), else (6), line (6), well (6), again (6), things (6), had (6), models (6), them (6), does (6), measurements (6), methods (6), groups (6), hie (6), their (6), speech (6), cheney (6), sets (5), interested (5), idea (5), see (5), left (5), combined (5), adjusted (5), gridding (5), represent (5), where (5), red (5), most (5), processing (5), grid (5), earth (5), true (5), anoms (5), way (5), calculated (5), going (5), will (5), therefore (5), factors (5), co2 (5), assumption (5), come (5), exactly (5), could (5), smoothing (5), records (5), rep (5), 250 (5), name (5), lat (5), 2008 (5), february (5), news (5), iraq (5), july (5), august (5), november (5), know (4), won (4), compared (4), such (4), example (4), included (4), plots (4), note (4), readings (4), degree (4), getting (4), measurement (4), trends (4), different (4), approach (4), average (4), significant (4), various (4), together (4), least (4), adjustment (4), function (4), frame (4), temps (4), par (4), col (4), lwd (4), substa (4), load (4), sciencapades (4), may (4), came (4), yearly (4), trying (4), gets (4), differences (4), apparently (4), back (4), should (4), any (4), due (4), 100 (4), computer (4), didn (4), down (4), often (4), sunday (4), bias (4), steve (4), done (4), thought (4), weather (4), without (4), lmer (4), altitude (4), algorithm (4), including (4), std (4), fm2 (4), being (4), think (4), heart (4), democrats (4), democracy (4), 2011 (4), palin (4), she (4), hate (4), march (4), response (3), days (3), change (3), possible (3), became (3), striking (3), good (3), individual (3), number (3), another (3), comparing (3), compare (3), averaging (3), case (3), concerned (3), ones (3), lines (3), step (3), effort (3), whole (3), necessary (3), sum (3), calculate (3), annual (3), less (3), paste (3), tab (3), complicated (3), result (3), seeing (3), because (3), friday (3), makes (3), sense (3), clear (3), just (3), close (3), process (3), perhaps (3), many (3), mcintyre (3), saturday (3), value (3), wanted (3), completely (3), lose (3), take (3), looked (3), ipcc (3), clouds (3), cloud (3), cover (3), decade (3), air (3), show (3), increases (3), thing (3), work (3), problem (3), others (3), last (3), question (3), find (3), mixed (3), within (3), seems (3), match (3), longer (3), following (3), long (3), tmp (3), memory (3), fm1 (3), suburban (3), variance (3), formula (3), estimates (3), squared (3), 500 (3), 1000 (3), gisstemp (3), government (3), better (3), got (3), grouping (3), his (3), power (3), group (3), linear (3), 127 (3), error (3), update (3), own (3), selected (3), still (3), designated (3), center (3), surrounding (3), thus (3), cia (3), growth (3), cities (3), detect (3), marked (3), every (3), published (3), union (3), republican (3), correct (3), sarah (3), whatever (3), totalitarianism (3), against (3), elite (3), 2012 (3), transplant (3), september (3), december (3), probably (2), houston (2), regard (2), especially (2), cause (2), land (2), pretty (2), doubt (2), reasons (2), hard (2), figure (2), period (2), final (2), overall (2), whether (2), warmer (2), mainly (2), valid (2), minded (2), very (2), gives (2), determining (2), blue (2), fall (2), squares (2), few (2), looks (2), comparison (2), code (2), extr1 (2), return (2), anomaly (2), minus (2), tapply (2), type (2), xlab (2), ylab (2), main (2), abline (2), coef (2), lty (2), ann (2), color (2), 2009 (2), sep (2), shocking (2), recall (2), post (2), calculating (2), point (2), trouble (2), top (2), once (2), interest (2), material (2), isn (2), stick (2), straightforward (2), similar (2), always (2), need (2), seen (2), big (2), cases (2), represented (2), during (2), faith (2), predictions (2), report (2), 2001 (2), modelers (2), resulting (2), tropics (2), atmosphere (2), around (2), fashion (2), until (2), evidence (2), started (2), only (2), accurate (2), really (2), determined (2), closer (2), shows (2), rate (2), order (2), per (2), disagreement (2), efforts (2), adjusting (2), among (2), introduce (2), proper (2), truth (2), matter (2), estimation (2), understood (2), modeling (2), seemingly (2), combination (2), tried (2), taken (2), lot (2), includes (2), dset2 (2), areas (2), increased (2), working (2), package (2), extr (2), use (2), running (2), bands (2), words (2), 750 (2), 1250 (2), 1500 (2), 1750 (2), 2000 (2), 2250 (2), 2500 (2), 2750 (2), 3000 (2), 3250 (2), 3500 (2), 3750 (2), 4000 (2), 4250 (2), 4500 (2), 4750 (2), 5000 (2), 5750 (2), book (2), enough (2), anything (2), occupy (2), amount (2), available (2), coverage (2), business (2), latitude (2), summary (2), fit (2), reml (2), aic (2), bic (2), loglik (2), deviance (2), remldev (2), dev (2), residual (2), obs (2), 548500 (2), 13483 (2), fixed (2), estimate (2), almost (2), shown (2), tuesday (2), prove (2), designations (2), obvious (2), previously (2), trend (2), list (2), database (2), thermometers (2), located (2), observed (2), countryside (2), urbanization (2), world (2), decline (2), log (2), axis (2), extent (2), reading (2), above (2), away (2), com (2), new (2), york (2), short (2), smaller (2), governor (2), sign (2), wisconsin (2), state (2), dictatorship (2), unions (2), atmospheric (2), predicted (2), caused (2), spot (2), her (2), read (2), bush (2), via (2), weiner (2), party (2), nothing (2), organizations (2), critical (2), people (2), who (2), islam (2), free (2), takes (2), identity (2), possession (2), books (2), rage (2), him (2), intelligence (2), weapons (2), nie (2), permits (2), explosion (2), october (2), sextant (2), view, complete, profile, boffin, involved, neuroscience, medical, service, email, frew, swbell, net, texas, united, states, subscribe, atom, home, older, except, oneself, downloaded, http, www, climateaudit, loaded, files, modern, flip, finds, segements, ignored, colored, produce, third, 1960, 1985, examples, 1890, 1975, upward, produced, justifies, important, planet, itself, affect, dispensed, voila, collected, painstakingly, speak, regression, part, meaning, determine, measurments, spends, considerable, making, reorganize, simpler, table, include, future, put, debugging, split, unsplit, set1, throw, green, brown, coral, darkgreen, orange, pink, sub, xlim, 1880, print, getgraphicsevent, onkeybd, keybd, old, share, discovery, evening, subtracting, describes, dashed, small, appears, unnecessary, terms, players, manner, great, deal, basic, massive, computerized, ought, something, far, locally, respectively, minimal, absolute, across, plotted, end, build, gnarly, box, basis, cancelling, quite, calculation, smooths, considerably, happens, removed, examined, dealing, difficulties, handled, iii, someone, told, took, anthropogenic, claims, face, crude, divides, 8000, continental, rain, guess, changes, consider, negate, doubling, natural, variability, larger, percent, slight, magnified, also, formation, offset, feedback, cycles, forcing, rinse, repeat, distribution, uniform, open, oceans, humidity, varies, constantly, pictures, swirls, assumptions, host, fudge, tweaked, endlessly, coughs, desired, agreement, validity, having, recently, reflected, agree, appreciate, care, clever, statistics, hell, three, fourths, globe, explain, limited, goes, four, alarm, according, noaa, sedate, 1972, rise, reconcile, wrong, jim, burning, guys, worked, postings, writings, months, ago, pointed, job, eliminating, algorithms, rather, astounding, cells, apply, series, steps, geometrical, introduction, sort, likely, longest, start, decreased, shorter, aren, usually, cement, lead, reorganized, information, folded, interp, substr, monthly, lme4, meter, contain, dumped, groupings, 999, your, 23716400, 04312743, 28028908, 0821297, 5952519, 2160810, 4049196, 4774994, 4415439, 2931860, 4352673, 7974976, 5588156, 6522236, 5811832, 2383055, 6373919, 7024097, 8671371, 2833311, 0728011, 3266110, 8495323, 0044707, 4843320, numbers, slope, coefficients, nearly, identical, 00035, sure, statistician, application, variation, improve, fidelity, averages, went, command, spacial, vindicated, respect, regardless, byzantine, proof, pudding, lack, guessing, reason, radiosonde, cheap, clearance, priced, desktop, duplicated, size, block, multi, analyzing, perform, impossible, factor, interesting, consistent, 1274202, 1274280, 637094, 1274190, 1274188, 510624, 71458, 113, 129532, 63624, 141931, 37674, 856394, 46410, 068179, 26111, 521113, 72188, 4514, 159, 1288347, 1288426, 644167, 1288337, 1288333, 67732, 41721, 14162, 37633, 330, 71739, 18564, 58698, 21184, 84184, 91752, 52133, 72203, 544, 427, 451, 6009074, 7884377, 2316191, 1136268, 5794809, 9054092, 3426933, 0768528, 4410112, 0043036, 0247369, 5251244, 8542720, 6135464, 0035847, 3222669, 1468110, 1510483, 3572983, 9318326, 5925205, 9627467, 6483137, 89551186, 03941906, 93493366, 0765255, 6208086, 5843329, 8087508, 5837928, 6445832, 4677458, 9302840, 6332613, 5261542, 1896612, 2221295, 5283583, 4532507, 3289638, 4701555, 5784181, 7600821, 5108772, 8686941, 7051168, 5278219, 0410194, 2047241, 3644670, 2930061, 4509212, 3098546, 3064547, 0101544, 3917199, 3916299, 7265894, 7631131, project, brilliantly, science, wednesday, continuing, mining, conclusion, robust, dominating, inaccurate, review, 300, sites, reveal, single, misclassification, parsed, combed, best, aggregate, necessarily, explanations, guesses, contaminated, sitting, asphalt, conditioner, compressors, purported, usa, northeast, flat, declining, 1950, brightness, cross, section, adequately, situation, enthusiasm, possibility, errors, emerging, worth, airport, versus, rest, described, hoped, hiding, scientist, seeking, claim, gradually, detailed, unable, demonstrate, nasa, hypothesis, address, occasion, country, proxy, correlation, region, figures, fact, assumed, exponential, depending, exp, decreases, proportional, attempt, mimick, plugging, relationship, randomly, countries, emerges, demonstrated, satisfaction, strange, arcane, cru, remains, affecting, begun, exist, yesterday, local, seem, inside, loop, suburbs, beyond, though, quickly, little, wunderderground, enables, dozens, simultaneously, chicago, regions, middle, manhattan, feet, unless, actually, maps, hit, vast, majority, needed, suspicion, too, urbanized, opposite, largest, centers, qualify, sprawl, longitude, direction, provided, nearest, tenth, below, story, added, comes, listed, disappears, unbiased, lost, briffa, finnish, tree, ring, reanalysis, common, theme, demonstrators, madison, today, walker, dictator, showed, after, sheds, light, public, members, particular, liberals, general, voters, elected, overwhelmingly, legislature, precisely, budget, under, control, senate, fled, masse, prevented, quorum, vote, couldn, place, frustrate, broadly, organized, labor, movement, priority, eliminate, secret, ballot, certification, elections, levels, bigger, missing, core, uses, unit, total, units, source, disagreements, misunderstandings, hot, troposphere, decades, balloons, satellites, failed, trace, incorrect, mid, 90s, aqua, plug, reasonable, surprise, happened, haven, scientists, familiar, acknowledged, suspect, funding, simply, disaster, scenario, field, sink, sight, obscurity, emails, document, dump, contemplates, birth, weeks, child, downs, syndrome, maturity, wisdom, equanimity, thoughtfulness, approaches, joy, trig, letter, monday, confounding, enemies, genius, shear, luck, instapundit, intense, eagerness, unfolded, concurrently, attention, paid, anthony, antics, agony, poor, journalists, must, endured, wienerisms, hurt, democratic, stood, ready, damage, misdeeds, awful, urrrgghhh, damn, family, values, frustration, becoming, popular, west, australia, discussing, implementation, media, council, ministry, reprimand, irresponsible, american, forever, advocating, forms, restriction, tolerance, term, expanded, form, canada, bureau, orwellian, human, rights, commission, prosecuting, run, afoul, dicates, certain, ethnic, sympathetic, muslim, insistance, criticized, self, censor, protect, christianity, religions, impression, regularity, regimentation, totalitarian, society, commentators, thomas, friedman, times, wax, eloquent, praise, chinese, communist, obstructions, thrown, nation, magazine, publish, stories, wing, mobs, buildings, disrupt, services, campus, venue, shut, intimidation, violence, leftist, act, bunch, brownshirts, never, dawn, concerns, passions, crushed, regime, imagine, otherwise, become, stay, realize, hands, exclude, cambodia, markers, glasses, decided, killing, fields, dick, eruption, hatred, our, ever, civil, cousins, funny, major, charges, turned, false, congressional, commissions, cook, lie, untoward, influence, decision, saying, everyone, clinton, knew, chemical, biological, restart, program, nuclear, normal, channels, vetted, director, usual, executive, wish, moreover, responsible, putting, track, energy, independence, gaining, despite, obama, administration, limit, oil, production, denying, federal, slow, rolling, offshore, blocking, keystone, pipeline, yet, guts, reality, based, 2002, april, 2003, 2004, 2005, 2006, blog, archive, notes, politics, issues, day, irritable, fool,

Text of the page (random words):
eds light i think on how public union members in particular and liberals in general think what is going on here is that the voters of wisconsin have elected a republican governor and overwhelmingly a republican legislature precisely so that they can get the state s budget under control what the democrats don t like isn t dictatorship it is democracy that is why the democrats in the wisconsin senate fled the state en masse they prevented a quorum so that a vote they were going to lose couldn t take place once again it is democracy they are trying to frustrate not dictatorship one could make the point more broadly about the organized labor movement the unions top priority is to eliminate the secret ballot in union certification elections why we all know why posted by rod at 5 43 pm saturday january 16 2010 briffa finnish tree ring reanalysis posted by rod at 8 44 pm sunday january 10 2010 the heat island effect lost and found as i ve demonstrated to my own satisfaction at least the strange arcane adjustments for hie made by cru and giss are not even necessary one gets the same trends whether they are made giss data set 2 or not giss data set 1 if an unbiased and well understood method for determining temperature anomalies is used what exactly they are doing with those adjustments is up in the air but the question remains to what extent the hie is affecting the results that we are seeing for global temperatures analysis of the giss station data had begun to make me think that the heat island effect did not exist but yesterday i was looking over station data from the local houston area and it did seem that there was an obvious hie individual stations in the heart of the city inside the loop were reading about 3 or 4 degrees above those out in the suburbs and beyond it was striking to me though how quickly the readings came back down to match those of the surrounding countryside getting even a little away from the city center looking at other cities i found the same thing wunderderground com enables one to look at dozens of individual stations in an area simultaneously note how in chicago the temperature readings decline as one gets away from the city center and here is the new york city area compare to the surrounding regions note the two stations in the middle of manhattan they are located 500 feet up unless stations were actually within the city they would not detect the effect so looking back over the maps of giss station locations it hit me that the vast majority of those stations marked as urban are not urban at all at least not to the extent that was seemingly needed to find a hie all this time the suspicion has been that the stations marked rural are too urbanized when it could be exactly the opposite those marked urban are often not that urban thus i selected stations from the 85 largest urban centers in the world in order the qualify the station had to be within the urban area urban sprawl for 0 1 degrees latitude or longitude in every direction from the designated station location which is provided to the nearest tenth of a degree the temperature anomalies for these as determined by the statistical algorithm used in previous posts is compared to the published giss data below to make a long story short as more cities are added in to the list of city stations included in the analysis above from smaller and smaller cities the trend line comes down to be much closer to that of the giss data and this is still well short only about 10 of the stations listed as urban in the giss dset1 data as i ve shown previously if all the urban stations are included the hie disappears posted by rod at 9 33 am saturday january 09 2010 population density and global warming it has often been observed that the center of a city is warmer than the surrounding countryside thus the claim has been made that global warming is an error due to weather stations gradually being taken over by urbanization as detailed in previous posts i ve been unable to demonstrate that from the weather station data available from nasa the hypothesis i address on this occasion is that idea that population density in a country could be used as a proxy of urbanization if so then there should be a correlation between population density change over time and the temperature in that region i used population density figures from the cia world fact book and assumed an exponential decline in density from that time depending on time d d 100 g 100 exp year 2008 127 where d population density in a given year d is the population density in 2008 g is the growth in population in 2008 year is the year of measurement thus d decreases in a log linear fashion over time proportional to the rate of growth observed in 2008 an attempt to mimick population growth plugging this group into the random effects statistical algorithm used in previous posts i found no relationship between d and temperature in 60 randomly selected countries the x axis is the log of d population density and the y axis is degrees c no effect of population density over time emerges with this analysis posted by rod at 4 01 pm friday january 08 2010 seeking a heat island effect continuing on with the data mining of the gisstemp material i have been looking for evidence of a heat island effect and my conclusion is that it won t be found in this data since my own analysis didn t show a robust hie i looked for bias in the dset1 data i thought it might be due to urban records being longer than rural ones and therefore dominating the data set this did not prove out the average length of rural and urban records is about the same i thought it might be inaccurate rural suburban and urban designations but a review of 300 rural sites selected at random did not reveal a single obvious misclassification no matter how i parsed and combed through this data i was getting no significant difference in rural and urban stations the best i ve done is to see a 1 8 degree difference in the aggregate analysis that i ve posted previously and that does not necessarily represent a trend so what are the possible explanations here is my list of guesses 1 the database is contaminated with thermometers that are sitting on asphalt or next to air conditioner compressors that are purported to be in rural locations 2 the population density in the areas of the usa where most of the thermometers are located the northeast has been flat or declining since 1950 therefore a grouping of u r or s or brightness in one cross section in the database does not adequately represent the situation i don t have lot of enthusiasm for working through this possibility it seems to me that even if the u r and s designations in the giss data had errors that there would still be some differences emerging for what it s worth here are the results of comparing locations that are designated rural and are not designated as airport locations versus all the rest of the locations using the random effects statistical algorithm described in previous posts these are not the results i hoped for but hiding them would make me no better than a climate scientist posted by rod at 8 15 pm wednesday january 06 2010 here s the problem here s the problem with the gisstemp station data not enough coverage posted by rod at 8 54 pm tuesday january 05 2010 climate science as steve mcintyre has so brilliantly shown jim hansen s methods for processing station temperatures into global temperatures don t really make sense the burning question for me therefore has been if the stick guys methods don t make sense and perhaps introduce bias then what do we get when proper methods are used what exactly is the truth of the matter and so i ve worked through steve s postings and the writings of others trying to find out what a proper method might be in a post from 18 months ago steve pointed out that the job of processing this data and eliminating the effects of various factors to come up with an estimation of the temperature anomalies over time could be done using well understood mixed effects statistical modeling algorithms i thought it rather astounding that a statistical method seemingly completely different from hansen s methods would come up with much the same results steve used this in modeling the combination of location data into grid cells but why not apply it to the whole data set and so i tried that giss dset0 you may recall is temperature data taken from different weather stations in some cases these stations are all together in one location as in 4 or 5 stations in a given city area hansen first combined these stations together for a location then he did a lot of complicated smoothing and adjustment of the data in a series of steps dset0 is raw data dset1 includes some adjustments and combination of data of stations at given locations dset2 includes more adjustments within a given location and gridding of the data does more of the same smoothing data over geometrical areas between locations with all of the adjusting and smoothing going on the introduction of some sort of bias seems likely for example hansen always used the longest records to start with and then increased or decreased the shorter records to match the longer but aren t the longer records usually from urban stations would this not cement the heat island effect into the combined location data and so working in the r project statistical package and following mcintyre s lead i first reorganized the data in giss dset0 into a simple long data frame with location information folded in from giss info extr function giss dset0 giss info year 0 temp 0 alt 0 ur 0 loc 0 lxs 0 sta 0 ni 1 length giss dset0 is na giss dset0 for i in ni if sum is na giss dset0 i 0 next else for j in 1 length giss dset0 i if is na giss dset0 i j next else n length giss dset0 i j 1 sta c sta rep names giss dset0 i j n year c year giss dset0 i j 1 temp c temp giss dset0 i j 18 ur c ur rep giss info urban i n alt c alt rep 250 giss info alt interp i 250 n tmp substr names giss dset0 i j 1 8 loc c loc rep tmp n lxs c lxs rep paste tmp i names giss dset0 i j sep n sta sta 1 year year 1 temp temp 1 alt alt 1 loc loc 1 ur ur 1 lxs lxs 1 return data frame loc loc sta sta year year ur ur alt alt lxs lxs temp temp dset0 t extr giss dset0 giss info no i could not use monthly data without running out of memory then calculate random effects load the lme4 package dset0 fm lmer temp 1 loc 1 sta 1 ur 1 alt 1 year data dset0 t dset0 fm1 lmer temp 1 sta 1 year data dset0 t loc location id sta station id with location ur urban vs suburban vs rural alt location altitude in 250 meter bands ranef dset0 fm year and ranef dset0 fm1 year contain the random effects year over year the temperature anomaly for each year in other words most of the variance is dumped into other groups using more groupings in the model formula makes no difference in these estimates dset0 fm vs dset0 fm1 r squared 0 999 here is your heat island effect ranef dset0 fm ur intercept r 0 23716400 rural s 0 04312743 suburban u 0 28028908 urban and here is the effect of altitude ranef dset0 fm alt intercept 0 10 0821297 250 9 5952519 500 8 2160810 750 7 4049196 1000 6 4774994 1250 5 4415439 1500 4 2931860 1750 3 4352673 2000 1 7974976 2250 0 5588156 2500 1 6522236 2750 5 5811832 3000 5 2383055 3250 13 6373919 3500 8 7024097 3750 13 8671371 4000 1 2833311 4250 6 0728011 4500 3 3266110 4750 0 8495323 5000 1 0044707 5750 0 4843320 what that yearly random effect looks like compared to the gisstemp book numbers and here is the difference between the two r squared 0 73 close enough for government work the slope coefficients are nearly identical a 0 00035 per year difference and i m sure that a statistician would do much better much of what hansen did in step 1 and 2 and in gridding the data was smoothing application of a smoothing algorithm would no doubt take out the variation in the random effects estimation i got but that would not improve the fidelity of the results dset0 annual averages are as close to the raw data as i can get without running out of computer memory i went from that data to a final result with one command in r if i can come up with the same results with no spacial factors at all other than altitude and station grouping then does gridding mean anything can the analysis represent anything more than the 25 of the earth s surface the stations occupy hansen is vindicated in a big respect this shows that his methods regardless how byzantine introduce no significant amount of bias where temperature trends are concerned with regard to the station data available the proof is in the pudding and all that that lack of coverage of the earth s surface in station data i m guessing is the reason for the difference in satellite radiosonde and station estimates of global temperature the amount of memory and processing power i have in my cheap clearance priced desktop pc would have been duplicated by a computer the size of a city block if at all back in the days hansen started doing this climate business his multi step process of analyzing station data was perhaps an effort to perform an analysis that would have been impossible using the statistical methods i used here update including another grouping factor latitude in 5 degree bands has no effect on the temperature anomalies but it s interesting to look at the random effect for that group dset0 fm2 lmer temp 1 ur 1 lat 1 alt 1 sta 1 year data dset0 t ranef dset0 fm2 lat intercept 90 48 0765255 85 30 6208086 80 30 5843329 75 26 8087508 70 17 5837928 65 10 6445832 55 0 4677458 50 1 9302840 45 4 6332613 40 7 5261542 35 10 1896612 30 7 2221295 25 16 5283583 20 17 4532507 15 18 3289638 10 18 4701555 5 18 5784181 0 18 7600821 5 19 5108772 10 19 8686941 15 19 7051168 20 17 5278219 25 15 0410194 30 11 2047241 35 7 3644670 40 3 2930061 45 0 4509212 50 0 3098546 55 5 3064547 60 9 0101544 65 13 3917199 70 16 3916299 75 20 7265894 80 22 7631131 and including that group gives more consistent altitude and urban estimates ranef dset0 fm2 alt intercept 1000 7 6009074 0 8 7884377 250 8 2316191 500 5 1136268 750 9 5794809 1000 5 9054092 1250 5 3426933 1500 4 0768528 1750 3 4410112 2000 2 0043036 2250 1 0247369 2500 0 5251244 2750 1 8542720 3000 0 6135464 3250 8 0035847 3500 3 3222669 3750 5 1468110 4000 5 1510483 4250 8 3572983 4500 8 9318326 4750 5 5925205 5000 5 9627467 5750 7 6483137 ranef dset0 fm2 ur intercept r 0 89551186 s 0 03941906 u 0 93493366 here is the summary from lmer linear mixed model fit by reml formula temp 1 loc 1 sta 1 ur 1 alt 1 year data dset0 t aic bic loglik deviance remldev 1274202 1274280 637094 1274190 1274188 random effects groups name variance std dev sta intercept 0 510624 0 71458 loc intercept 113 129532 10 63624 year intercept 0 141931 0 37674 alt intercept 29 856394 5 46410 ur intercept 0 068179 0 26111 residual 0 521113 0 72188 number of obs 548500 groups sta 13483 loc 4514 year 127 alt 23 ur 3 fixed effects estimate std error t value intercept 5 ...