726 Matching Annotations
  1. Sep 2025
  2. Aug 2025
    1. This paper describes the principles behind the method of ComparativeJudgement, and in particular Adaptive Comparative Judgement, a technique borrowed frompsychophysics which is able to generate extremely reliable results for educationalassessment, and which is based on the kind of holistic evaluation that we assume was thebasis for judgement in pre-marking days, and that the users of assessment results expectour assessment schemes to capture

      comparative adaptive judgement, better than shoe-horning into fixed pre-conceived rubrics

  3. Jul 2025
  4. Apr 2025
  5. Mar 2025
  6. Feb 2025
    1. EOF3 describes a distinctive east/west gradient

      Altitude gradient, not just east-west! This is our vapor lakes mode I hypothesize, but with its contribution to the regional mean (EOF1) and a N-S gradient (EOF2) subtracted off, because the big dogs eat first in maximum covariance analyses like EOFs.

    1. fferent solutions for tropical deep convection, depending on what boundary layer scheme is used

      Miura figure from early days of NICAM high-res global modeling

  7. Jan 2025
  8. earthobservatory.nasa.gov earthobservatory.nasa.gov
  9. Dec 2024
  10. d197for5662m48.cloudfront.net d197for5662m48.cloudfront.net
    1. show strongly suppressed cold pooldensity currents versus a Control simulation. Yet we find no statistically significant difference between the diurnal cycles ofMCS counts, with continued nocturnal propagation of storms

      Innteresting, are they coupled waves rather than surface-driven squalls?

  11. Sep 2024
    1. according to which it is only the mass convergence aloft that has an effect on the hydrostatic surface pressure

      Wallace and Hobbs Eq. 7.41 is just not right

  12. Aug 2024
  13. Feb 2024
    1. apparently nonstationary visualbehavior arises because of the non-Gaussian statistic

      anyway the point is: simple 'red' noise can have surprisingly large or long-lasting excursions, in fact random walks diverge, the drunk will never find that lamp again....

    1. The interactions can result in predator–prey cycles in moisture mode activity and Hadley cell strength that are akin to ITCZ breakdown.

      like the old "index cycles" of the jet stream

  14. Jan 2024
  15. Dec 2023
  16. Nov 2023
    1. Defininations: heatwave: at least three consecutive days in which the maximum 2-meter air temperature exceeds the 90th percentiles.  The consecutive days of heatwave is counted within a month, which is the same in other variables, such as CDD, CSDI, CWD, LCS, and LWS.   cold spell: the minimum 2-meter air temperature be less than the 10th percentiles.   warm spell: the maximum 2-meter air temperature exceeds the 90th percentiles.    percentiles: in this collection they are based on daily temperature and precipitation fields, and calculated using a running window of +/- 7 days centered on each day of the year for the climatology period of 1991-2020 in Version 2 (1981-2010 in Version 1).

      Definitions: a key aspect

    1. percentage of time when daily max 2-m temperature < 10th percentile

      1990-2020 TIME average guaranteed to be exactly 10% everywhere. Should we check that, for some calendar month?

    1. The monthly percentiles indicate atwhich percentile the monthly mean falls relative to the 1991-2020 period.

      monthly or daily same 30-year reference period

    2. ercentiles were calculated using a running window of +/- 7 days centered on each day ofthe year for the climatology period of 1991 through 2020 for version 2 products.

      MERRA2 statistics products, percentiles definition and reference period (1991-2020)

  17. Oct 2023
    1. evolving lowpass filtered fields act as a "steering flow” for the high pass filtered perturbations

      just temporal filtering, but it suffices to separate long vs. short waves in spatial scale

    2. strength of the condensation heating in the model can be controlled by specifying the relative humidity

      development takes a little longer with weaker latent heating

    3. those shown in Part II move in tandem with features embedded in the base of the updraft, indicative of a deeper structure

      0:27 has nice deep layered wav e clouds above the surface gust front

    4. aerosols are categorized by tint: dust (orange), smoke (red) organic/black carbon (green), and sulfates (white)

      long lovely one, the once a day data assimilation window for tracers sweeps westward it seems.

    1. CONDA_SUBDIR=osx-64 conda create

      prepend conda with this for new Apple M1/M2 chips if ordinary conda -c conda-forge fails

      Successful: CONDA_SUBDIR=osx-64 conda create -n MetPy_arm64 -c conda-forge metpy siphon xarray jupyterlab

  18. Jun 2023
  19. May 2023
  20. Mar 2023
  21. Jan 2023
  22. Nov 2022
    1. Up-scale growth thus occurs as a continuing amplification oflarge scales after small scales saturate, which begs the question of what sets the shape of the saturation(climatological) power spectra. Wind spectra are nearly power-law with a logarithmic slope of about −5/3in the free troposphere, remarkably so in the 2D runs and clearly distinct from slope −2 (a null hypoth-esis of spectrally white wind divergence). A common interpretation of −5/3 slope − as indicative of a cas-cade, a steady conservative transfer of energy from source to sink scales by interactions that are local inlog-wavelength space − is hard to apply to these moist tropical waves.

      predictability aspects of convectively coupled waves. When the cascade example is elevated to paradigm, it cuts off the view to wider pastures -- not that I have a counter-theory, just a sense that a question can be asked starkly

  23. Oct 2022
  24. Sep 2022
  25. Aug 2022
    1. that more knowledge about climateintervention approaches and their consequences will help society make informed, justdecisions about the deployment of climate intervention. AGU is not taking a position aboutspecific climate interventions, but AGU is making the case that a robust body of scientificevidence about climate intervention and an ethical framework should be available

      The grammatical singular word "society" is quite problematic here, in my view. Likewise, the premise of 'support research first, with ethics relegated to adjectives like responsible in a framework' is a problematic overarching framework. Perhaps to be expected from a scientific society, a research = funding = activity = influence maximizing enterprise, but as likely in my view to be a cause as a preventer of ethically good outcomes overall.

  26. May 2022
  27. Apr 2022
  28. Mar 2022
    1. The dominant deep convective heating process changes from latent heating at low levels to eddy heat-flux convergence in the upper troposphere. This requires a substantial updraught-environment temperature difference

      This is the fundamental reason that convective buoyancy is plentiful is upper levels. Because the SIGN of buopyancy is crucial, but the value not so important, low-level environmental density (which can make near-zero lifted-parcel buoyancy be slightly positive or negative) is more important than density (T) at upper levels in modulating convection in tropical waves.

  29. Nov 2021
  30. weather.rsmas.miami.edu weather.rsmas.miami.edu
  31. Oct 2021
  32. weather.rsmas.miami.edu weather.rsmas.miami.edu
    1. The dynamics of many mesoscale phenomena are oftenmore intuitive if the vertical momentum equation is writtenin terms of a buoyancy force and a vertical perturbationpressure gradient force.

      Here is the basic definition of buoyancy

  33. Sep 2021
    1. Widths of the earliest detectable low-level updrafts associated with sustained precipitating deep convection were ~3–5 km, larger than updrafts associated with surrounding boundary layer turbulence (~1–3 km wide)

      Bigger updrafts are better

    1. close agreement with Fig. 1.11.

      The jet stream is in thermal wind balance with a latitudinal T gradient. The coriolis force holds up a wall of cold air, so gravity can’t pull it down

    1. 25

      J. A. Dutton,The Ceaseless Wind. An Introduction to the Theory of AtmosphericMotion(McGraw-Hill, New York, NY, 1976)

      Not exactly "extensive"!

      KE "decomposition" isn't right, there is a cross term

      Energetics (cross term as interaction) seems more promising

      Using vorticity budgets through time to infer the divergent flow, using $$d/dt(\zeta) = \zeta \times \delta$$ are called the "chi problem" (Sardeshmukh 1993).

    2. dividing the circulation of a terrestrial tidallylocked atmosphere into rotational and divergent componentscorresponds to a division into two physically meaningful circu-lations. The rotational circulation is primarily composed of thestationary Rossby waves forced by the divergent circulation andthe zonal-mean jet produced by the stationary part of the atmo-spheric circulation.

      Summary

    Annotators

    1. The energetics of this system describe exchanges among the divergent and the rotational energy components. The energy equations are as follows.

      Energetics of divergent-rotational interactions

    1. Boyle, James S. October 1996, 34 pp.  The 200 hPa kinetic energy is represented by means of the spherical harmonic components for the AMIP simulations, the NCEP/NCAR reanalysis and the ECMWF reanalysis(ERA). The data used are the monthly mean wind fields for 1979 to 1988. The kinetic energy is decomposed into the divergent (DKE) and rotational (RKE) components and emphasis is placed on examining the former.

      but they are not orthogonal, so this is not a true decomposition of KE

  34. Aug 2021
  35. Jun 2021
    1. f the major premise is made sensible by making itprobabilistic, not absolute, the syllogism becomes for-mally incorrect and leads to a conclusion that is not sen-sible

      Null hypothesis significance testing is problematic

  36. May 2021
    1. Norbert Wiener, a child prodigy and a great mathematician, coined the term 'cybernetics' to characterize a very general science of 'control and communication in the animal and machine'. It brought together concepts from engineering, the study of the nervous system and statistical mechanics (e.g. entropy). From these he developed concepts that have become pervasive through science (especially biology and computing) and common parlance: 'in-formation', 'message', 'feedback' and 'control'. He wrote, 'the thought of every age is reflected in its technique . . . If the seventeenth and early eighteenth centuries are the age of clocks, and the later eighteenth and nineteenth centuries constitute the age of steam engines, the present time is the age of communication and control.'

      He coined "feedback"?? I can't imagine a scientific language without that term. Profound.

    1. where slow and shigh are respectively the saturated moist entropy averaged over the altitude ranges of 1–3 km and 5–7 km

      II defined 550-430 mb is 5-7km, 890-700 is 1-3km

    2. where is threshold entropy, defined as the average of the saturated moist entropy in the 1.5–2 km layer, minus the boundary layer moist entropy sbl averaged over the 0–1 km layer

      DCIN defined

    1. he MPO Ph.D. degree requires a minimum of 30 course credits, of which a minimum of 9 course credits should be taken from 700 level courses.

      30 COURSE credits, at least 9 from 700 level classes

      total credits are 60

  37. Apr 2021
  38. Mar 2021
    1. A peculiar feature in MERRA-2 is that its moisture analysis increment (estimated by budget residual) is positive and largest even though CWV anomalies are larger than those in the other RAs

      MERRA-2 facts

    2. The contribution of LW to MSE propagation in MERRA-2 and JRA-55 is relatively small, which is likely related to the relatively weak cloud–radiation feedback strength

      MERRA-2 facts

  39. Feb 2021
    1. LRP highlights the regionsof greater relevance for the ANN to predict the year by propagating an output sample back-ward through the frozen nodes of the ANN until it reaches the input layer

      positive-only map? Why is the sign lost?

    1. virtual sv(red broken), moist h

      The virtual effect of vapor (lowering air density by its lighter molecular weight) involves (0.61q)T, about 0.2 K for a 1 g/kg vapor increment. If that same 1g/kg of vapor's latent heat is released (so that it lowers the air's density by warming), it increases T by (\(L/C_p\)) or 2.5C. This factor of about (2.5)/0.2 = 12.5 can be seen graphically here at low levels, comparing \(s_v-s\) to \(h-s\).

      By the way, if that 1 g/kg of condensed water is carried by the parcel (falling at its terminal velocity), its contribution to buoyancy is g/1000, the density equivalent of cooling the air by about 0.3K.

    1. It is not currently possible to “un-flag” an annotation — if an annotation is flagged by mistake, the group creator can choose not to hide it.

      Wow, this is so emblematic of the age... any person can taint another irrevocably with a flag!

      And an accident cannot be undone!

      Now when I screen share on Zoom to discuss our document, everyone will see a big red flag.

    Tags

    Annotators

    URL

  40. Jan 2021
    1. a damaged species: associate professors. Long forgotten, these tenured professors find themselves burdened with extensive service and administrative tasks

      Hmm, I thought it was the other way around: more tedious forms of service (because of rules requiring full professor status) awaited after promotion, so I lingered a long time as Associate.

  41. Nov 2020
  42. Oct 2020
    1. Advζ, AdvT, and the Q term, respectively,

      These are called QGO_PVA (positie vorticity advection above), QGO_WA (warm advection), and QGO_dia (for diabatic). The sum QGO_PVA + QGO_WA = QGO_adia, for adiabatic, the traditiional view of winter QG omega.

    2. The total ω may be separated into a component explained by the QG ω equation (ωqg) and the remainder [ωaqg, calculated as a residual term in Eq. (2)]. A significant advantage of the QG ω equation is its linearity, and thus, we may directly decompose ωqg into components forced by the large-scale advective forcings and diabatic heating.

      All those derivatives (and their inverse) are a linear operation, op(A + B + C) = op(A) + op(B) + op(C)

    1. there will alwaysremain some degree of uncertainty (or errors) inthe initial conditions and, due to the nonlinearityof atmospheric motions, these errors inevitablyamplify with time. Beyond some threshold forecastinterval the forecast fields are, on average, no morelike the observed fields against which they are veri-fied than two randomly chosen observed fields forthe same time of year are like one another. For theextratropical atmosphere this so-called limit ofdeterministic predictabilityis believed to be on theorder of 2 weeks

      predictability limit: fundamental, inescapable

    2. It can be shownthat, in the integral over the entire mass of theatmosphere, the generation of kinetic energy by theV term in (7.42) is equal to the release ofpotential energy associated with the rising of warmair and sinking of cold air.

      perhaps in flux form...\(\overline{\alpha \omega}\) is the term

  43. Sep 2020
    1. 0

      Zero, PLUS a curl-of-friction term (viscous effects), PLUS a vertical derivative of heating term. These are fundamental drivers and dampers that make otherwise-conservation have interesting situations rather than 0=0 forever.

    1. standard value used in our simulations is the “pure mixing time scale” of 3,600 s

      LLCS is relaxation toward an adjusted (neutralized) target.

      The target is defined by an upward process of 100% mixing of adjacent layers between lowest unstable parcel and the top level where the crest of this upward-mixing pile of air, lifted one more level, is no longer warmer than ambient.

    2. Hence, instead of truncating, we rotate β and γ back into θ,q on levels by forming the 2m component vector and the 2m × 2m matrix to interpret our results

      Inputs are rotated back into profile for plotting

    3. the vast majority of output convective behavior can be described with relatively few eigenvectors. We typically retain two or three for discussion in the results sections and use 10 when the emulators are run online as part of the GCM

      output truncation is here

    4. We predict Q from P rather than predicting Y from X because correlations between values of θ and q on different vertical levels that could cause large errors in our statistical analysis are avoided

      Orthogonal basis, but not truncated. Maybe the small late-in-the-list modes are no problem to bring along for completeness, since their spurious (sample-specific) correlations are down-weighted by their smallness. Truncation occurs below, actually.

    5. Cases from the equally spaced group deemed nonconvecting (cpΔθtrop < 0.05 MJ m−2) are then discarded, as are an equal number of nonconvecting cases. This leaves an equal number of convecting and nonconvecting cases

      10K/d is the threshold

    6. The choice of 60,000 cases was made, as it is realistic to perform analysis on matrices of this size with available computing resources. It is found that using smaller sample sizes has little effect on accuracy

      matrix software is great

    7. nonconvecting cases (defined as cpΔθtrop <0.05 MJ m−2, equivalent to a temperature change ΔTtrop  ≃ 10.1 K day−1,

      10K/d heating of 700-100mb layer is not "no convection"!!

    8. the cases closest to 30,000 equally spaced values of cpΔθtrop from its minimum to its maximum value

      No averaging, just sampling to an equally spaced grid in the space of range

    9. A two‐step linear statistical emulator is used

      Linear closure, but with a trigger. But trigger is really just a screening of their training dataset for places where the emulated schemes themselves have convection disabled (trigger off).

    10. The overall aim is not to replace conventional parameterization nor to improve GCM integration speed but to understand our parameterizations in the context of process knowledge and provide tools for parameterization development and interpretation of climate model projections.

      Right, just making sense of the situation -- speed is no good if direction is unclear.

    11. The derived response matrices must also be put in the GCM in place of the original parameterization, as was done by Kelly et al. (2017) and Mapes et al. (2019)

      Yay, our dream

    1. spatial organization of tropical deep convection, can manifest itselfin two ways: through a decrease in the total area covered by convection and/or through a decrease in thenumber of convective areas

      two kinds of organization

    1. David (1997)−2.0543Soulik (2006)−1.7317Susan (1988)−2.0144Mamie (1985)−1.6818Dale (1996)−2.0045Phanfone (2002)−1.6719Thelma (1980)−2.0046Agnes (1981)−1.6720Lola (1986)−1.9747Dianmu (2004)−1.67

      Team 4 do these

    2. Polly (1995)−1.9248Ann (1999)−1.6622Dan (1995)−1.8949Podul (2001)−1.6623Nabi (2005)−1.8850Ma-on (2004)−1.6524Dinah (1987)−1.8751Krosa (2001)−1.6325Faxai (2007)−1.8752Winnie (1997)−1.63

      Team 5 do these

    3. Xangsane (2000)−2.0838Forrest (1986)−1.7512Stella (1998)−2.0839Kinna (1991)−1.7513Choi-wan (2009)−2.0840Irma (1981)−1.7414Vanessa (1984)−2.0641Marge (1983)−1.7315Elsie (1992)−2.0642Gay (1981)−1.73

      Team 3 do these

    4. Orchid (1980)−2.3133Cecil (1993)−1.767Nanmadol (2004)−2.1834Judy (1982)−1.768Rammasun (2008)−2.1735Irving (1979)−1.769Mireille (1991)−2.1536Owen (1979)−1.7510Songda (2004)−2.1337Flo (1990)−1.75

      Team 2 do these

    5. 1Melor (2009)−2.5828Chaba (2004)−1.832Oscar (1995)−2.5729Colleen (1989)−1.823Hunt (1992)−2.3630Mac (1982)−1.794Haishen (2002)−2.3531Gordon (1982)−1.785Elsie (1981)−2.3232Megi (2004)−1.77

      Team 1 do these

    6. Schematic representation of upper-tropospheric ridge amplification and jet streak intensification associated with divergent TC outflow impinging upon an upper-tropospheric jet stream

      Archambault schematic of ET extratropical transition mechanism of Rossby wave forcing