Validation and calibration of models to estimate photosynthetically active radiation considering different time scales and sky conditions
Photosynthetically Active Radiation (PAR) is a fundamental parameter for developing plant productivity models. Nevertheless, instrumentation for measuring PAR and to record it is scarce at conventional meteorological stations. Several procedures have therefore been proposed for PAR estimation. In th...
| Autores: | , , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | España |
| Recursos: | Universidad Pública de Navarra |
| Repositorio: | Academica-e. Repositorio Institucional de la Universidad Pública de Navarra |
| OAI Identifier: | oai:academica-e.unavarra.es:2454/43733 |
| Acesso em linha: | https://hdl.handle.net/2454/43733 |
| Access Level: | acceso abierto |
| Palavra-chave: | Photosynthetically active radiation Solar irradiance Sky types Model calibration |
| Resumo: | Photosynthetically Active Radiation (PAR) is a fundamental parameter for developing plant productivity models. Nevertheless, instrumentation for measuring PAR and to record it is scarce at conventional meteorological stations. Several procedures have therefore been proposed for PAR estimation. In this work, 21 previously published analytical models that correlate PAR with easily available meteorological parameters are collected. Although longer time scales were considered in the original publications, a minute range was applied in this work to calibrate the PAR models. In total, more than 10 million input records were gathered from the SURFRAD station network from a 10-year long time series with data frequencies recorded every 1 min. The models were calibrated both globally, using data from all stations and locally, with data from each station. After calibration, the models were validated for minute, hourly and daily data, obtaining low fitting errors at the different stations in all cases, both when using the globally calibrated models and with the models calibrated for each location. Although the PAR results in general improved for locally calibrated models, the use of local models is not justified, since the global models presented offered very satisfactory PAR results for the different climatic conditions where the meteorological stations are located. Thus, PAR estimation model should then be selected, solely considering the meteorological variables available at the specific location. When applying the globally calibrated models to input data classified according to sky conditions (from clear to overcast), the PAR models continued to perform satisfactorily, although the error statistics of some models for overcast skies worsened. |
|---|