Abstract

Accurate estimation of dew point temperature (T<sub>dew</sub>) plays a very important role in the fields of water resource management, agricultural engineering, climatology and energy utilization. However, there are few studies on the applicability of local T<sub>dew</sub> algorithms at regional scales. This study evaluated the performance of a new machine learning algorithm, i.e., gradient boosting on decision trees with categorical features support (CatBoost) to estimate daily T<sub>dew</sub> using limited local and cross-station meteorological data. The random forests (RF) algorithm was also assessed for comparison. Daily meteorological data from 2016 to 2019, including maximum, minimum and average temperature (T<sub>max</sub>, T<sub>min</sub> and T<sub>mean</sub>), maximum, minimum and average relative humidity (RH<sub>max</sub>, RH<sub>min</sub> and RH<sub>mean</sub>), maximum, minimum and average global solar radiation (Rs<sub>max</sub>, Rs<sub>min</sub> and Rs<sub>mean</sub>) from three weather stations in Hunan of China were used to evaluate the CatBoost and RF algorithms. The results showed that both algorithms achieved satisfactory estimation accuracy at the target stations (on average RMSE = 1.020°C, R<sup>2 </sup>=<sup> </sup>0.969, MAE = 0.718°C and NRMSE = 0.087) in the absence of complete meteorological parameters (with only temperature data as input). The CatBoost algorithm (on average RMSE = 1.900°C and R<sup>2 </sup>=<sup> </sup>0.835) was better than the RF algorithm (on average RMSE = 2.214°C and R<sup>2 </sup>=<sup> </sup>0.828). The accuracy and stability of the CatBoost and RF algorithms were positively correlated with the number of input parameters, and the three-parameter algorithms achieved higher estimation accuracy than the two-parameter algorithms. The developed methodology is helpful to predict T<sub>dew</sub> at regional scale.

Highlights

  • Dew point temperature (Tdew) is the temperature at which water vapor in the air condenses into water droplets

  • We evaluate the applicability of random forests (RF) and CatBoost algorithms for estimating Tdew under the local input scenario, using meteorological data from Fenghuang, Huayuan and Longshan stations in China

  • This paper evaluated the applicability of a new algorithm (CatBoost) under two input scenarios combined with limited meteorological data from different regional stations in China to accurately estimate daily Tdew and extend it to regional applications

Read more

Summary

Introduction

Dew point temperature (Tdew) is the temperature at which water vapor in the air condenses into water droplets. Accurate estimation of Tdew data plays a significant role in the fields of energy utilization [1], thermal energy [2,3] and engineering [4]. Tdew is usually used in conjunction with relative humidity to calculate the water content in the air [7]. It can be combined with the wet bulb. Tdew is needed to estimate reference crop evapotranspiration (ET0) [9]. Compared with other meteorological variables, Tdew is still relatively inadequate. Because of its importance and non-linear changes, accurate estimation of Tdew has vital scientific significance in the above fields

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call