Rooftop dew, fog and rain collection in southwest Morocco and predictive dew modeling using neural networks

详细信息	查看全文 | 推荐本文 |

• 作者：Imad Lekouch^a ; ^b ; ^c ; Khalid Lekouch^a ; Marc Muselli^c ; ^d ; Anne Mongruel^b ; Belkacem Kabbachi^a ; Daniel Beysens^b ; ^c ; ^e ; ^{daniel.beysens@espci.fr}
• 关键词：Dew formation ; Artificial neural network ; Water production ; Radiative condenser ; Hydrometeorology
• 刊名：Journal of Hydrology
• 出版年：2012
• 期刊代码：68_00221694
• 类别：abs
• 出版时间：2 July, 2012
• 卷：448-449
• 期：Complete
• 页码：60-72
• 文件大小：2149 K

摘要

| Figures/TablesFigures/Tables | ReferencesReferences

Summary

Two coastal sites were investigated in an arid region of southwest Morocco to determine the amount of dew, fog and rain that could be collected from rooftops for household use. Systematic measurements were performed in Mirleft (43 m asl, 200 m from the coast) for 1 year (May 1, 2007 to April 30, 2008) and in Id Ouasskssou (240 m asl, 8 km from the coast) for three summer months (July 1, 2007 to September 30, 2007). Dew water was collected using standard passive dew condensers and fog water by utilizing planar fog collectors. The wind flow was simulated on the rooftop to establish the location of the fog collector. At both sites, dew yields and, to a lesser extent, fog water yields, were found to be significant in comparison to rain events. Mirleft had 178 dew events (48.6%of the year, 18 卤 2 L m^鈭? cumulated amount) and 20 fog episodes (5.5%of the year, 1.4 L m^鈭? with uncertainty 鈭?.2/+0.4 L m^鈭? cumulated amount), corresponding to almost 40%of the yearly rain contribution (31 rain events, 8.5%of the year, 49 卤 7 mm cumulated amount). At Id Ouasskssou there were 50 dew events (7.1 卤 0.3 L m^鈭?, 54.3%frequency), 16 fog events (6.5 L m^鈭? with uncertainty 鈭?.1/+1.8 L m^鈭?, 17.4%frequency) and six rain events (16 卤 2 mm, 6.5%frequency).

Meteorological data (air and dew point temperature and/or relative humidity, wind speed and wind direction, cloud cover) were recorded continuously at Mirleft to assess the influence of local meteorological conditions on dew and fog formation. Using the set of collected data, a new model for dew yield prediction based on artificial neural networks was developed and tested for the Mirleft site. This model was then extrapolated to 15 major cities in Morocco to assess their potential for dew water collection. It was found that the location of the cities with respect to the Atlas mountain chain, which controls the circulation of the humid marine air, is the main factor that influences dew production.