Acta Optica Sinica, Volume. 43, Issue 6, 0601002(2023)

Cloud Base Height Retrieval Methods for FY-4A Based on Ensemble Learning

Zhuofu Yu1, Ya Wang2、*, Shuo Ma1、**, Weihua Ai1, and Wei Yan1
Author Affiliations
  • 1College of Meteorology and Oceanography, National University of Defense Technology, Changsha 410000, Hunan, China
  • 2National Satellite Meteorological Center, China Meteorological Administration, Beijing 100081, China
  • show less
    Figures & Tables(19)
    Flow chart of data matching between FY-4A and CloudSat
    Schemes of CBH retrieval for FY-4A designed in this paper
    Variation of RMSE of eight types of clouds with decision trees number in RF model and GBT model. (a) Variation of RMSE with decision trees number in RF model; (b) variation of RMSE with decision trees number in GBT model
    Variation of RMSE of eight types of clouds with maximum depth of decision trees in RF model and GBT model. (a) Variation of RMSE with maximum depth of decision trees in RF model; (b) variation of RMSE with maximum depth of decision trees in GBT model
    Variation of RMSE of samples on training and validation datasets with decision trees number in RF model and GBT model. (a) Variation of RMSE with decision trees number in RF model; (b) variation of RMSE with decision trees number in GBT model
    Variation of RMSE of samples on training and validation datasets with maximum depth of decision trees in RF model and GBT model. (a) Variation of RMSE with maximum depth of decision trees in RF model; (b) variation of RMSE with maximum depth of decision trees in GBT model
    Retrieval results of models of two schemes on test dataset. (a) Scheme one; (b) scheme two
    Retrieval flow of CBH for FY-4A in practical application
    CBH retrieved from models of two schemes and the comparison with CBH from CloudSat. (a) Cloud types obtained according by the model proposed in Ref. [32]; (b) CBH retrieved from the cloud types of Fig. 9(a) and the model of scheme one; (c) CBH retrieved from the model of scheme two; (d) comparison between the cloud types of CloudSat and the model proposed in Ref. [32], and comparison among CBH retrieved from models of two schemes and CBH from CloudSat on CloudSat track
    • Table 1. Parameters of FY-4A/AGRI channels

      View table

      Table 1. Parameters of FY-4A/AGRI channels

      Channel numberCenter wavelength /μmSpatial resolution /km
      10.471.0
      20.650.5
      30.8251.0
      41.3752.0
      51.612.0
      62.252.0
      73.75(H)2.0
      83.75(L)4.0
      96.254.0
      107.14.0
      118.54.0
      1210.74.0
      1312.04.0
      1413.54.0
    • Table 2. Related cloud products of FY-4A/AGRI and corresponding channels

      View table

      Table 2. Related cloud products of FY-4A/AGRI and corresponding channels

      Cloud product of FY-4A/AGRICenter wavelength of optical channel /μm
      Cloud top height(CTH)11.0,12.09,13.55
      Cloud optical thickness(COT)and cloud effective radius(CER)0.65,2.25
    • Table 3. Numbers of all types of clouds after data matching

      View table

      Table 3. Numbers of all types of clouds after data matching

      Cloud typeNumber
      Ci73255
      As71000
      Ac41412
      St/Sc104792
      Cu24535
      Ns40397
      Dc10719
      Multi65603
    • Table 4. Decision trees number of two models for all types of clouds

      View table

      Table 4. Decision trees number of two models for all types of clouds

      Cloud typeDecision trees number of RF modelDecision trees number of GBT model
      Ci1393
      As971
      Ac1383
      St/Sc117
      Cu137
      Ns137
      Dc53
      Multi1127
    • Table 5. Maximum depth of decision trees of two models for all types of clouds

      View table

      Table 5. Maximum depth of decision trees of two models for all types of clouds

      Cloud typeMaximum depth in RFMaximum depth in GBT
      Ci85
      As85
      Ac74
      St/Sc96
      Cu45
      Ns88
      Dc56
      Multi85
    • Table 6. Retrieval results of two models for all types of clouds on the test dataset

      View table

      Table 6. Retrieval results of two models for all types of clouds on the test dataset

      Cloud

      Mean

      value /m

      RF modelGBT model
      RMSE /mMAE /mRMRE /%RMSE /mMAE /mRMRE /%
      Ci9238.51060.6841.40.72499.471054.8836.50.72859.41
      As4918.61524.51180.90.648141.821515.81173.70.653241.65
      Ac2954.91357.21063.70.639666.681352.21060.00.643166.51
      St/Sc973.2620.0372.50.606243.45669.5394.40.592146.06
      Cu1075.61000.3634.40.501878.631035.0663.50.513284.22
      Ns1266.41338.8896.20.4246106.351371.7933.20.4210113.18
      Dc712.9534.5369.50.184760.12538.4375.60.183161.73
      Multi2285.02044.71593.70.4046148.922044.91605.50.4069150.72
    • Table 7. Optimal CBH retrieval model of all types of clouds

      View table

      Table 7. Optimal CBH retrieval model of all types of clouds

      Cloud typeRetrieval model
      CiGBT
      AsGBT
      AcGBT
      St/ScRF
      CuRF
      NsRF
      DcRF
      MultiRF
    • Table 8. Retrieval results of RF model and GBT model on test dataset

      View table

      Table 8. Retrieval results of RF model and GBT model on test dataset

      ModelRMSE /mMAE /mRMRE /%
      RF2113.41497.60.7769124.60
      GBT2109.11498.60.7779124.81
    • Table 9. Retrieval results of RF model and GBT model for all types of clouds on test dataset

      View table

      Table 9. Retrieval results of RF model and GBT model for all types of clouds on test dataset

      Cloud

      type

      Mean value /mRF modelGBT model
      RMSE /mMAE /mRMRE /%RMSE /mMAE /mRMRE /%
      Ci9238.52528.72178.00.648923.802529.72181.50.651823.84
      As4918.61896.31532.00.611141.191891.81529.50.614041.04
      Ac2954.91637.81246.10.556663.071636.41247.30.556663.17
      St/Sc973.21098.2536.60.278266.401095.6537.70.276366.69
      Cu1075.61734.91080.00.2450171.411725.91076.70.2458170.77
      Ns1266.41980.81518.90.2353246.691977.51527.60.2320248.25
      Dc712.93090.62524.6-0.0628545.123071.92532.3-0.0578546.44
      Multi2285.03104.52369.90.3775297.833095.72367.40.3783297.88
    • Table 10. Retrieval results of models of two schemes for all types of clouds on test dataset

      View table

      Table 10. Retrieval results of models of two schemes for all types of clouds on test dataset

      CloudMean value /mScheme oneScheme two
      RMSE /mMAE /mRMRE /%RMSE /mMAE /mRMRE /%
      Ci9238.51054.8836.50.72859.412529.72181.50.651823.84
      As4918.61515.81173.70.653241.651891.81529.50.614041.04
      Ac2954.91352.21060.00.643166.511636.41247.30.556663.17
      St/Sc973.2620.0372.50.606243.451095.6537.70.276366.69
      Cu1075.61000.3634.40.501878.631725.91076.70.2458170.77
      Ns1266.41338.8896.20.4246106.351977.51527.60.2320248.25
      Dc712.9534.5369.50.184760.123071.92532.3-0.0578546.44
      Multi2285.02044.71593.70.4046148.923095.72367.40.3783297.88
    Tools

    Get Citation

    Copy Citation Text

    Zhuofu Yu, Ya Wang, Shuo Ma, Weihua Ai, Wei Yan. Cloud Base Height Retrieval Methods for FY-4A Based on Ensemble Learning[J]. Acta Optica Sinica, 2023, 43(6): 0601002

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Atmospheric Optics and Oceanic Optics

    Received: Apr. 13, 2022

    Accepted: Aug. 1, 2022

    Published Online: Mar. 13, 2023

    The Author Email: Wang Ya (ywang@cma.gov.cn), Ma Shuo (mashuo0601@163.com)

    DOI:10.3788/AOS220957

    Topics