Technical note: Using distributed temperature sensing for Bowen ratio evaporation measurements

Abstract

Abstract. Rapid improvements in the precision and spatial resolution of distributed temperature sensing (DTS) technology now allow its use in hydrological and atmospheric sciences. Introduced by Euser et al. (2014) is the use of DTS for measuring the Bowen ratio (BR-DTS), to estimate the sensible and latent heat flux. The Bowen ratio is derived from DTS-measured vertical profiles of the air temperature and wet-bulb temperature. However, in previous research the measured temperatures were not validated, and the cables were not shielded from solar radiation. Additionally, the BR-DTS method has not been tested above a forest before, where temperature gradients are small and energy storage in the air column becomes important.In this paper the accuracy of the wet-bulb and air temperature measurements of the DTS are verified, and the resulting Bowen ratio and heat fluxes are compared to eddy covariance data. The performance of BR-DTS was tested on a 46 m high tower in a mixed forest in the centre of the Netherlands in August 2016. The average tree height is 26 to 30 m, and the temperatures are measured below, in, and above the canopy. Using the vertical temperature profiles the storage of latent and sensible heat in the air column was calculated.We found a significant effect of solar radiation on the temperature measurements, leading to a deviation of up to 3 K. By installing screens, the error caused by sunlight is reduced to under 1 K. Wind speed seems to have a minimal effect on the measured wet-bulb temperature, both below and above the canopy. After a simple quality control, the Bowen ratio measured by DTS correlates well with eddy covariance (EC) estimates (r 2 = 0.59). The average energy balance closure between BR-DTS and EC is good, with a mean underestimation of 3.4 W m −2 by the BR-DTS method. However, during daytime the BR-DTS method overestimates the available energy, and during night-time the BR-DTS method estimates the available energy to be more negative. This difference could be related to the biomass heat storage, which is neglected in this study.The BR-DTS method overestimates the latent heat flux on average by 18.7 W m −2 , with RMSE = 90 W m −2 . The sensible heat flux is underestimated on average by 10.6 W m −2 , with RMSE = 76 W m −2 . Estimates of the BR-DTS can be improved once the uncertainties in the energy balance are reduced. However, applying, for example, Monin-Obukhov similarity theory could provide independent estimates for the sensible heat flux. This would make the determination of the highly uncertain and difficult to determine net available energy redundant.

Estimation of Temperature and Associated Uncertainty from Fiber-Optic Raman-Spectrum Distributed Temperature Sensing

“…The application of DTS is of great value to characterize thermal dynamics at a scale that corresponds to, for example, many geophysical processes. Heat tracer tests with DTS have been used to estimate wind speed [3,4], evaporation [5,6], soil moisture [7], groundwater-surface water interaction [1,8], and groundwater flow [9,10]. The specification sheets of DTS measurement systems list the uncertainty in the temperature estimated under ideal conditions.…”

Reviews and syntheses: Turning the challenges of partitioning ecosystem evaporation and transpiration into opportunities

“…T was subsequently approximated as the difference between measured ET and the model for E soil during times when photosynthesis was active. Annual T / ET values from this approach varied from 0.35 to 0.66 in the grass ecosystem (US-Dk1) across a 4-year period and between 0.7 and 0.75 in the pine (US-Dk3) and hardwood (US-Dk2) forests, somewhat higher than global syntheses (Schlesinger and Jasechko, 2014), remote sensing estimates from PT-JPL (see Appendix A) for the Duke pine forest (Fig. 3), and sap-flow-based measurements from the deciduous forest (Oishi et al, 2008).…”

“…(Wei et al, 2017). Other observational studies suggest that annual T / ET averages nearly 2/3 globally (0.61 ± 0.15, Schlesinger and Jasechko, 2014;0.64 ± 0.13, Good et al, 2015; and 0.66 ± 0.13 across some FLUXNET sites, Li et al, 2019). Intercomparison studies agree on the large uncertainty surrounding these estimates, with reported global terrestrial annual T / ET ratios ranging from 0.35 to 0.90 (Coenders-Gerrits et al, 2014;Fatichi and Pappas, 2017;Young-Robertson et al, 2018).…”

“…The ratio of transpiration to evapotranspiration (T / ET) at annual timescales is related to aridity (Good et al, 2017) but appears to be relatively insensitive to annual precipitation (P ) (Schlesinger and Jasechko, 2014). T / ET is sensitive to ecosystem characteristics, namely the leaf area index (LAI) Fatichi and Pappas, 2017;Wang et al, 2014;Wei et al, 2015), especially on sub-annual timescales (Li et al, 2019;Scott and Biederman, 2017), noting that LAI is related to P at longer timescales.…”

“…The mean monthly latent heat flux (λE) -the energy used for evapotranspiration -from eddy covariance measurements from four research sites ("MEASURED") and 13 ecosystem models from the North American Carbon Program Site-Level Interim Synthesis (Schwalm et al, 2010). Sites: CA-Ca1 (Schwalm et al, 2007), CA-Obs (Griffis et al, 2003;Jarvis et al, 1997), US-Ho1 (Hollinger et al, 1999), US-Me2 (Thomas et al, 2009). Models: BEPS (Liu et al, 1999), CAN-IBIS (Williamson et al, 2008), CNCLASS (Arain et al, 2006), ECOSYS (Grant et al, 2005), ED2 (Medvigy et al, 2009), ISAM (Jain and Yang, 2005), ISOLSM (Riley et al, 2002), LOTEC (Hanson et al, 2004), ORCHIDEE (Krinner et al, 2005), SIB , SIBCASA (Schaefer et al, 2009), SSIB2 (Zhan et al, 2003), TECO (Weng and Luo, 2008).…”