The retrieval of accurate profiles of temperature and water vapour is importantfor the study of atmospheric convection. Recent development in computationaltechniques motivated us to use adaptive techniques in the retrieval algorithms.In this work, we have used an adaptive neuro-fuzzy inference system (ANFIS) toretrieve profiles of temperature and humidity up to 10 km over the tropicalstation Gadanki (13.5° N, 79.2° E), India. ANFIS istrained by using observations of temperature and humidity measurements byco-located Meisei GPS radiosonde (henceforth referred to as radiosonde) andmicrowave brightness temperatures observed by radiometrics multichannelmicrowave radiometer MP3000 (MWR). ANFIS is trained by considering theseobservations during rainy and non-rainy days (ANFIS(RD + NRD)) and during non-rainy days only (ANFIS(NRD)). The comparison of ANFIS(RD + NRD) andANFIS(NRD) profiles with independent radiosonde observations and profilesretrieved using multivariate linear regression (MVLR: RD + NRD and NRD)and artificial neural network (ANN) indicated that the errors in theANFIS(RD + NRD) are less compared to other retrieval methods.The Pearson product movement correlation coefficient () between retrievedand observed profiles is more than 92% for temperature profiles for alltechniques and more than 99% for the ANFIS(RD + NRD) techniqueTherefore this new techniques is relatively better for the retrieval oftemperature profiles. The comparison of bias, mean absolute error (MAE), RMSE and symmetric mean absolute percentage error(SMAPE) of retrieved temperature and relative humidity (RH) profiles using ANN and ANFIS alsoindicated that profiles retrieved using ANFIS(RD + NRD) are significantlybetter compared to the ANN technique. The analysis of profiles concludes thatretrieved profiles using ANFIS techniques have improved the temperatureretrievals substantially; however, the retrieval of RH by all techniquesconsidered in this paper (ANN, MVLR and ANFIS) has limited success.
展开▼
