The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning
<p>Mixed-phase Southern Ocean clouds are challenging to simulate, and their representation in climate models is an important control on climate sensitivity. In particular, the amount of supercooled water and frozen mass that they contain in the present climate is a predictor of their planetar...
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Copernicus Publications
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/5ec23559377445169f369c25275355b6 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| id |
oai:doaj.org-article:5ec23559377445169f369c25275355b6 |
|---|---|
| record_format |
dspace |
| institution |
DOAJ |
| collection |
DOAJ |
| language |
EN |
| topic |
Environmental engineering TA170-171 Earthwork. Foundations TA715-787 |
| spellingShingle |
Environmental engineering TA170-171 Earthwork. Foundations TA715-787 R. Atlas J. Mohrmann J. Finlon J. Lu I. Hsiao R. Wood M. Diao The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| description |
<p>Mixed-phase Southern Ocean clouds are challenging to simulate, and their
representation in climate models is an important control on climate
sensitivity. In particular, the amount of supercooled water and frozen mass
that they contain in the present climate is a predictor of their planetary
feedback in a warming climate. The recent Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES) vastly increased the
amount of in situ data available from mixed-phase Southern Ocean clouds useful
for model evaluation. Bulk measurements distinguishing liquid and ice water
content are not available from SOCRATES, so single-particle phase
classifications from the Two-Dimensional Stereo (2D-S) probe are invaluable
for quantifying mixed-phase cloud properties. Motivated by the presence of
large biases in existing phase discrimination algorithms, we develop a novel
technique for single-particle phase classification of binary 2D-S images using
a random forest algorithm, which we refer to as the University of Washington
Ice–Liquid Discriminator (UWILD). UWILD uses 14 parameters computed from
binary image data, as well as particle inter-arrival time, to predict phase.
We use liquid-only and ice-dominated time periods within the SOCRATES dataset
as training and testing data. This novel approach to model training avoids
major pitfalls associated with using manually labeled data, including reduced
model generalizability and high labor costs. We find that UWILD is well
calibrated and has an overall accuracy of 95 <span class="inline-formula">%</span> compared to
72 <span class="inline-formula">%</span> and 79 <span class="inline-formula">%</span> for two existing phase classification
algorithms that we compare it with. UWILD improves classifications of small
ice crystals and large liquid drops in particular and has more flexibility
than the other algorithms to identify both liquid-dominated and ice-dominated
regions within the SOCRATES dataset. UWILD misclassifies a small percentage
of large liquid drops as ice. Such misclassified particles are typically
associated with model confidence below 75 <span class="inline-formula">%</span> and can easily be
filtered out of the dataset. UWILD phase classifications show that particles
with area-equivalent diameter (<span class="inline-formula"><i>D</i><sub>eq</sub></span>) <span class="inline-formula"><</span> 0.17 <span class="inline-formula">mm</span> are mostly
liquid at all temperatures sampled, down to <span class="inline-formula">−</span>40 <span class="inline-formula"><sup>∘</sup>C</span>. Larger
particles (<span class="inline-formula"><i>D</i><sub>eq</sub>>0.17</span> <span class="inline-formula">mm</span>) are predominantly frozen at all
temperatures below 0 <span class="inline-formula"><sup>∘</sup>C</span>. Between 0 and 5 <span class="inline-formula"><sup>∘</sup>C</span>,
there are roughly equal numbers of frozen and liquid mid-sized particles (<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M14" display="inline" overflow="scroll" dspmath="mathml"><mrow><mn mathvariant="normal">0.17</mn><mo><</mo><msub><mi>D</mi><mtext>eq</mtext></msub><mo><</mo><mn mathvariant="normal">0.33</mn></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="87pt" height="14pt" class="svg-formula" dspmath="mathimg" md5hash="c7daab499ad5c37ae38e052da514de7f"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="amt-14-7079-2021-ie00001.svg" width="87pt" height="14pt" src="amt-14-7079-2021-ie00001.png"/></svg:svg></span></span> <span class="inline-formula">mm</span>), and larger particles (<span class="inline-formula"><i>D</i><sub>eq</sub>>0.33</span> <span class="inline-formula">mm</span>) are mostly frozen. We also use UWILD's phase
classifications to estimate sub-1 <span class="inline-formula">Hz</span> phase heterogeneity, and we show
examples of meter-scale cloud phase heterogeneity in the SOCRATES dataset.</p> |
| format |
article |
| author |
R. Atlas J. Mohrmann J. Finlon J. Lu I. Hsiao R. Wood M. Diao |
| author_facet |
R. Atlas J. Mohrmann J. Finlon J. Lu I. Hsiao R. Wood M. Diao |
| author_sort |
R. Atlas |
| title |
The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| title_short |
The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| title_full |
The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| title_fullStr |
The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| title_full_unstemmed |
The University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning |
| title_sort |
university of washington ice–liquid discriminator (uwild) improves single-particle phase classifications of hydrometeors within southern ocean clouds using machine learning |
| publisher |
Copernicus Publications |
| publishDate |
2021 |
| url |
https://doaj.org/article/5ec23559377445169f369c25275355b6 |
| work_keys_str_mv |
AT ratlas theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jmohrmann theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jfinlon theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jlu theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT ihsiao theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT rwood theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT mdiao theuniversityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT ratlas universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jmohrmann universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jfinlon universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT jlu universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT ihsiao universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT rwood universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning AT mdiao universityofwashingtoniceliquiddiscriminatoruwildimprovessingleparticlephaseclassificationsofhydrometeorswithinsouthernoceancloudsusingmachinelearning |
| _version_ |
1718439039749586944 |
| spelling |
oai:doaj.org-article:5ec23559377445169f369c25275355b62021-11-11T14:06:16ZThe University of Washington Ice–Liquid Discriminator (UWILD) improves single-particle phase classifications of hydrometeors within Southern Ocean clouds using machine learning10.5194/amt-14-7079-20211867-13811867-8548https://doaj.org/article/5ec23559377445169f369c25275355b62021-11-01T00:00:00Zhttps://amt.copernicus.org/articles/14/7079/2021/amt-14-7079-2021.pdfhttps://doaj.org/toc/1867-1381https://doaj.org/toc/1867-8548<p>Mixed-phase Southern Ocean clouds are challenging to simulate, and their representation in climate models is an important control on climate sensitivity. In particular, the amount of supercooled water and frozen mass that they contain in the present climate is a predictor of their planetary feedback in a warming climate. The recent Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES) vastly increased the amount of in situ data available from mixed-phase Southern Ocean clouds useful for model evaluation. Bulk measurements distinguishing liquid and ice water content are not available from SOCRATES, so single-particle phase classifications from the Two-Dimensional Stereo (2D-S) probe are invaluable for quantifying mixed-phase cloud properties. Motivated by the presence of large biases in existing phase discrimination algorithms, we develop a novel technique for single-particle phase classification of binary 2D-S images using a random forest algorithm, which we refer to as the University of Washington Ice–Liquid Discriminator (UWILD). UWILD uses 14 parameters computed from binary image data, as well as particle inter-arrival time, to predict phase. We use liquid-only and ice-dominated time periods within the SOCRATES dataset as training and testing data. This novel approach to model training avoids major pitfalls associated with using manually labeled data, including reduced model generalizability and high labor costs. We find that UWILD is well calibrated and has an overall accuracy of 95 <span class="inline-formula">%</span> compared to 72 <span class="inline-formula">%</span> and 79 <span class="inline-formula">%</span> for two existing phase classification algorithms that we compare it with. UWILD improves classifications of small ice crystals and large liquid drops in particular and has more flexibility than the other algorithms to identify both liquid-dominated and ice-dominated regions within the SOCRATES dataset. UWILD misclassifies a small percentage of large liquid drops as ice. Such misclassified particles are typically associated with model confidence below 75 <span class="inline-formula">%</span> and can easily be filtered out of the dataset. UWILD phase classifications show that particles with area-equivalent diameter (<span class="inline-formula"><i>D</i><sub>eq</sub></span>) <span class="inline-formula"><</span> 0.17 <span class="inline-formula">mm</span> are mostly liquid at all temperatures sampled, down to <span class="inline-formula">−</span>40 <span class="inline-formula"><sup>∘</sup>C</span>. Larger particles (<span class="inline-formula"><i>D</i><sub>eq</sub>>0.17</span> <span class="inline-formula">mm</span>) are predominantly frozen at all temperatures below 0 <span class="inline-formula"><sup>∘</sup>C</span>. Between 0 and 5 <span class="inline-formula"><sup>∘</sup>C</span>, there are roughly equal numbers of frozen and liquid mid-sized particles (<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M14" display="inline" overflow="scroll" dspmath="mathml"><mrow><mn mathvariant="normal">0.17</mn><mo><</mo><msub><mi>D</mi><mtext>eq</mtext></msub><mo><</mo><mn mathvariant="normal">0.33</mn></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="87pt" height="14pt" class="svg-formula" dspmath="mathimg" md5hash="c7daab499ad5c37ae38e052da514de7f"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="amt-14-7079-2021-ie00001.svg" width="87pt" height="14pt" src="amt-14-7079-2021-ie00001.png"/></svg:svg></span></span> <span class="inline-formula">mm</span>), and larger particles (<span class="inline-formula"><i>D</i><sub>eq</sub>>0.33</span> <span class="inline-formula">mm</span>) are mostly frozen. We also use UWILD's phase classifications to estimate sub-1 <span class="inline-formula">Hz</span> phase heterogeneity, and we show examples of meter-scale cloud phase heterogeneity in the SOCRATES dataset.</p>R. AtlasJ. MohrmannJ. FinlonJ. LuI. HsiaoR. WoodM. DiaoCopernicus PublicationsarticleEnvironmental engineeringTA170-171Earthwork. FoundationsTA715-787ENAtmospheric Measurement Techniques, Vol 14, Pp 7079-7101 (2021) |