A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data
This paper proposes a complete framework of a machine learning-based model that detects convective initiation (CI) from geostationary meteorological satellite data. The suggested framework consists of three main processes: (1) An automated sampling tool; (2) machine learning-based CI detection model...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/11/12/1454 |
| _version_ | 1798024401858330624 |
|---|---|
| author | Daehyeon Han Juhyun Lee Jungho Im Seongmun Sim Sanggyun Lee Hyangsun Han |
| author_facet | Daehyeon Han Juhyun Lee Jungho Im Seongmun Sim Sanggyun Lee Hyangsun Han |
| author_sort | Daehyeon Han |
| collection | DOAJ |
| description | This paper proposes a complete framework of a machine learning-based model that detects convective initiation (CI) from geostationary meteorological satellite data. The suggested framework consists of three main processes: (1) An automated sampling tool; (2) machine learning-based CI detection modelling; (3) repeated model tuning through validation. In this study, the automated sampling tool was able to track the CI objects iteratively, even without ancillary data such as an atmospheric motion vector (AMV). The collected samples were used to train the machine learning model for CI detection. Random forest (RF) was used to classify the CI and non-CI. To enhance the advantages of the machine learning approach, we adopted model tuning to iteratively update the training dataset from each validation result by adding hits and misses to the CI samples, and false alarms and correct negatives to the non-CI samples. Using 12 interest fields from the Himawari-8 Advanced Himawari Imager (AHI) over the Korean Peninsula, this simple and intuitive tuning process increased the overall probability of detection (POD) from 0.79 to 0.82 and decreased the overall false alarm rate (FAR) from 0.46 to 0.37 with around 40 min of the lead-time. Amongst the 12 interest fields, <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(11.2) µm was identified as the most significant predictor in the RF model, followed by <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(8.6—11.2) µm, and <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(6.2−7.3) µm. The effect of model tuning on the CI detection performance was also analyzed using spatiotemporal validation maps. By automatically collecting and updating the machine learning training dataset, the suggested framework is expected to help the maintenance of the CI detection model from an operational perspective. |
| first_indexed | 2024-04-11T18:01:49Z |
| format | Article |
| id | doaj.art-5151224488ee4ae2a993ae2bb7d36f74 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-04-11T18:01:49Z |
| publishDate | 2019-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5151224488ee4ae2a993ae2bb7d36f742022-12-22T04:10:26ZengMDPI AGRemote Sensing2072-42922019-06-011112145410.3390/rs11121454rs11121454A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite DataDaehyeon Han0Juhyun Lee1Jungho Im2Seongmun Sim3Sanggyun Lee4Hyangsun Han5School of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan 44919, KoreaSchool of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan 44919, KoreaSchool of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan 44919, KoreaSchool of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan 44919, KoreaCentre for Polar Observation and Monitoring, University College London, Gower Street, London WC1E 6BT, UKUnit of Arctic Sea Ice Prediction, Korea Polar Research Institute, Incheon 21990, KoreaThis paper proposes a complete framework of a machine learning-based model that detects convective initiation (CI) from geostationary meteorological satellite data. The suggested framework consists of three main processes: (1) An automated sampling tool; (2) machine learning-based CI detection modelling; (3) repeated model tuning through validation. In this study, the automated sampling tool was able to track the CI objects iteratively, even without ancillary data such as an atmospheric motion vector (AMV). The collected samples were used to train the machine learning model for CI detection. Random forest (RF) was used to classify the CI and non-CI. To enhance the advantages of the machine learning approach, we adopted model tuning to iteratively update the training dataset from each validation result by adding hits and misses to the CI samples, and false alarms and correct negatives to the non-CI samples. Using 12 interest fields from the Himawari-8 Advanced Himawari Imager (AHI) over the Korean Peninsula, this simple and intuitive tuning process increased the overall probability of detection (POD) from 0.79 to 0.82 and decreased the overall false alarm rate (FAR) from 0.46 to 0.37 with around 40 min of the lead-time. Amongst the 12 interest fields, <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(11.2) µm was identified as the most significant predictor in the RF model, followed by <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(8.6—11.2) µm, and <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>T</mi> <mi>b</mi> </msub> </mrow> </semantics> </math> </inline-formula>(6.2−7.3) µm. The effect of model tuning on the CI detection performance was also analyzed using spatiotemporal validation maps. By automatically collecting and updating the machine learning training dataset, the suggested framework is expected to help the maintenance of the CI detection model from an operational perspective.https://www.mdpi.com/2072-4292/11/12/1454convective initiation (CI)Himawari-8random forestautomated samplingmachine learningrepeated model tuning |
| spellingShingle | Daehyeon Han Juhyun Lee Jungho Im Seongmun Sim Sanggyun Lee Hyangsun Han A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data Remote Sensing convective initiation (CI) Himawari-8 random forest automated sampling machine learning repeated model tuning |
| title | A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data |
| title_full | A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data |
| title_fullStr | A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data |
| title_full_unstemmed | A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data |
| title_short | A Novel Framework of Detecting Convective Initiation Combining Automated Sampling, Machine Learning, and Repeated Model Tuning from Geostationary Satellite Data |
| title_sort | novel framework of detecting convective initiation combining automated sampling machine learning and repeated model tuning from geostationary satellite data |
| topic | convective initiation (CI) Himawari-8 random forest automated sampling machine learning repeated model tuning |
| url | https://www.mdpi.com/2072-4292/11/12/1454 |
| work_keys_str_mv | AT daehyeonhan anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT juhyunlee anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT junghoim anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT seongmunsim anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT sanggyunlee anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT hyangsunhan anovelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT daehyeonhan novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT juhyunlee novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT junghoim novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT seongmunsim novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT sanggyunlee novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata AT hyangsunhan novelframeworkofdetectingconvectiveinitiationcombiningautomatedsamplingmachinelearningandrepeatedmodeltuningfromgeostationarysatellitedata |