On the estimation of boundary layer heights: a machine learning approach
oleh: R. Krishnamurthy, R. K. Newsom, L. K. Berg, H. Xiao, P.-L. Ma, D. D. Turner
| Format: | Article |
|---|---|
| Diterbitkan: | Copernicus Publications 2021-06-01 |
Deskripsi
<p>The planetary boundary layer height (<span class="inline-formula"><i>z</i><sub><i>i</i></sub></span>) is a key parameter used in atmospheric models for estimating the exchange of heat, momentum, and moisture between the surface and the free troposphere. Near-surface atmospheric and subsurface properties (such as soil temperature, relative humidity, etc.) are known to have an impact on <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span>. Nevertheless, precise relationships between these surface properties and <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span> are less well known and not easily discernible from the multi-year dataset. Machine learning approaches, such as random forest (RF), which use a multi-regression framework, help to decipher some of the physical processes linking surface-based characteristics to <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span>. In this study, a 4-year dataset from 2016 to 2019 at the Southern Great Plains site is used to develop and test a machine learning framework for estimating <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span>. Parameters derived from Doppler lidars are used in combination with over 20 different surface meteorological measurements as inputs to a RF model. The model is trained using radiosonde-derived <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span> values spanning the period from 2016 through 2018 and then evaluated using data from 2019. Results from 2019 showed significantly better agreement with the radiosonde compared to estimates derived from a thresholding technique using Doppler lidars only. Noteworthy improvements in daytime <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span> estimates were observed using the RF model, with a 50 % improvement in mean absolute error and an <span class="inline-formula"><i>R</i><sup>2</sup></span> of greater than 85 % compared to the Tucker method <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span>. We also explore the effect of <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span> uncertainty on convective velocity scaling and present preliminary comparisons between the RF model and <span class="inline-formula"><i>z</i><sub><i>i</i></sub></span> estimates derived from atmospheric models.</p>