Long-term Observations of Energy Balance Fluxes Over a Tropical Residential Neighborhood and Application to Urban Microclimate Modeling

Detailed, multi-year radiation and energy fluxes for a residential neighborhood in a tropical humid and wet climate (viz Singapore) are presented. The measurements cover a period of ~7 years (2006-2013) and represent the longest set of flux data reported for a tropical city. Owing to the equatorial location of the site, a number of interesting seasonal features are observed, which differ from those found in the more commonly researched mid-latitude cities. Incoming radiation is uniformly high throughout the year, with values which are amongst the highest reported by any urban study. Annual changes in climate and surface-atmosphere energy exchange are therefore much less than observed in cities located outside the tropics. The traditional seasonal classification (i.e. summer vs winter season) does not apply in the tropical context. Rather, significant variability exists using a classification based on clouds and rainfall (i.e. wet vs dry season). The energy balance partitioning is nevertheless similar to that reported for subtropical and mid-latitude suburban sites. Across the entire study period and all weather conditions ~54% of net radiation is partitioned into sensible and ~39% into latent heat, respectively, resulting in a long-term daily Bowen ratio of ~1.4.

The present data can further be used to evaluate and adapt models to investigate the urban heat island, improve the thermal comfort of residents or study extremes of urban weather. The present results have subsequently been used to test a mesoscale model against this tropical dataset. The TEB-ISBA land surface modeling has been evaluated in off-line mode with a continuous one year long dataset of radiation, energy balance and selected surface and air temperature measurements. Results show underestimation of net radiation, overestimation of sensible heat fluxes and increased storage heat fluxes. Model performance is comparable to that of the better performing model cohort in the PILPS-Urban multi-model comparison project. However, large systematic errors are found for several variables. This indicates that model performance can be further improved. For example, the vegetation parameterization used is inadequate to represent the full moisture dynamics, producing unrealistically dry conditions during a particularly dry period.

Singapore provides a unique climatic context, and the present long-term study is expected to add robust statistics from the understudied (sub)tropical region to the global dataset of urban energy balance fluxes, which is dominated by work conducted in mid and high latitudes.