Design and Evaluation of a Low-Cost Horizontal Tube Radiation Shield for Temperature and Humidity Measurements in a Tropical Urban Environment

Junnaedhi, I Dewa G. A.; Arkan, Muhammad; Muzasyaroh, An Nur A.; Nugraha, Dwina; Fitriani, Rachmy; Abdillah, Muhammad R.; Suwarman, Rusmawan; Sukendra, Sukendra; Varquez, Alvin C. G.; Inagaki, Atsushi; Kanda, Manabu

doi:10.5194/egusphere-2026-4598

Preprints

https://doi.org/10.5194/egusphere-2026-4598

Preprints

24 Sep 2026

| 24 Sep 2026

Status: this preprint is open for discussion and under review for Atmospheric Measurement Techniques (AMT).

Design and Evaluation of a Low-Cost Horizontal Tube Radiation Shield for Temperature and Humidity Measurements in a Tropical Urban Environment

I Dewa G. A. Junnaedhi, Muhammad Arkan, An Nur A. Muzasyaroh, Dwina Nugraha, Rachmy Fitriani, Muhammad R. Abdillah, Rusmawan Suwarman, Sukendra Sukendra, Alvin C. G. Varquez, Atsushi Inagaki, and Manabu Kanda

Abstract. Climate change and rapid urbanization have increased the need for dense air temperature and humidity monitoring networks in tropical cities. Although low-cost sensors are widely available, the cost of radiation shields remains a major barrier to large-scale deployment. This study presents a low-cost horizontal tube radiation shield (H-Tube) constructed from commercially available pipes and fittings combined with a small number of 3D-printed components. Two versions were developed: a bright-annealed stainless-steel shield (H-Tube_SS) and a PVC shield covered with aluminum foil (H-Tube_PVC). Material costs were approximately USD 25 and USD 11, respectively, making them about 8 and 18 times less expensive than typical commercial radiation shields (>USD 200).

Performance was evaluated through comparative field experiments in a tropical urban environment in Bandung, Indonesia, using identical Vaisala HMP155A temperature-humidity probes. Measurements from the H-Tube were compared against a Cotton Region Shelter (CRS) reference and two commercial radiation shields: a multiplate (MP) shield and a vertical tube (V-Tube) shield. Under non-aspirated conditions, daytime temperature biases exceeded 4 °C and relative humidity biases exceeded 15% during periods of strong net radiation. In contrast, aspirated operation substantially improved performance, with temperature biases generally remaining within ±1 °C and relative humidity biases within ±5%, comparable to those of the commercial MP and V-Tube shields. The aspirated H-Tube_SS achieved a temperature RMSE of 0.49 °C relative to the CRS, comparable to the MP shield (0.53 °C) and slightly lower than the H-Tube_PVC (0.58 °C).

Simple bias-correction methods based on linear temperature regression and saturation-vapor-pressure adjustment for relative humidity were also developed and evaluated. For the H-Tube designs, temperature RMSE was reduced from 1.35-1.46 °C to approximately 0.56 °C under non-aspirated conditions and from 0.49-0.58 °C to 0.30-0.35 °C under aspirated conditions. Relative humidity RMSE was reduced to approximately 2.6-3.2% after correction. The correction methods also reduced temperature-measurement uncertainty, particularly for the non-aspirated configurations.

These results demonstrate that the aspirated H-Tube provides a practical, scalable, and low-cost alternative to commercial radiation shields for distributed temperature-humidity monitoring networks in tropical urban environments.

Received: 30 Jul 2026 – Discussion started: 24 Sep 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Status: open (until 28 Nov 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-4598', Anonymous Referee #1, 14 Oct 2026 reply

Overall comments:

The manuscript lays the foundation for creating a low-cost, horizontal radiation shield for use in urban environments when making Temperature/Relative Humidity measurements. Such shield designs are needed for specific applications and the authors do a good job of designing a reasonable shield that prioritizes their goals. The manuscript is well laid out and presents the findings well. I have a few technical comments related to the design itself that I think bear further consideration.
Major Comments:
1). If other shield designs exist that address some or all of the stated concerns for the tropical urban environment the authors are studying, why not use those? Are there deficiencies in the currently available designs that the authors can state that require designing a new shield? From reading the introduction, my guess would be that the authors are attempting to create a substantially cheaper alternative (admirable and perfectly acceptable goal), but this isn’t clearly stated. I think adding a bit of detail to the introduction that highlights why this need exists could be useful. If the other shields that are available work for example, but are cost prohibitive, that needs to be clearly stated.
2). The double wall design is excellent and practical given the solar heating concerns, but is relatively narrow compared to the size of the HMP155 sensor as the authors state towards the end of the manuscript. This is particularly relevant for the HMP155 as depending on the filter used, this could significantly restrict airflow inside the shield and to the sensor itself. Were there tests done to explore the aspiration rate inside the inner tube or variations on overall size to preserve airflow?
3). The 45 degree angle inlet could potentially allow for wind driven rain to be pushed inside the horizontal section, has this been tested during particularly windy events? I see a wire mesh on the inlet and it’s stated there is one on the outlet. This could restrict airflow further, and for the inlet may provide a space for water to congregate, leading to excessive wetbulb errors
4). Reflected radiation on the inlet could lead to heating of the wire mesh, increasing temperature errors. Upstream thermal mass could modify temperature readings. This may come out more as an error source depending on the environment it is used it, more testing would be required to identify this, but do the authors have any comments regarding whether the wire mesh itself is providing any source of thermal mass?
5). Space for a single sensor does not allow for second faster temperature measurement, leading to slower response time given the time lag of the HMP. During periods of rapid change this will lead to excessive temperature errors and humidity errors. Did the authors consider making a larger diameter version that could accommodate multiple sensors?
6). It is unclear where the cable for the HMP comes out and how this could affect the airflow. Either text or an illustration should be added to make this more clear.
Minor comments:
Line 71-72 – it is true that multiplate shields are sensitive to wind-driven rain, but this statement should have a few references to go with it.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-4598-RC1
- AC1: 'Reply on RC1', Junnaedhi I Dewa Gede Agung, 23 Oct 2026 reply
  
  Dear Reviewer,
  Thank you very much for your positive assessment of our work. Indeed, the primary goal of this study is to develop an affordable temperature and humidity sensor network for tropical urban environments in Indonesia. We are also grateful for your constructive comments on the manuscript. Many of the issues you raised were concerns that we ourselves considered during the development and evaluation of the radiation shield.
  To support our responses to several of your comments, we have included a separate document entitled Response Figures. Please find the document attached to this response.
  For clarity, our responses are organized according to the structure of your comments. Thank you again for the time and effort you devoted to evaluating our manuscript and providing valuable feedback.
  Sincerely,
  
  Authors
  
  Major Comments:
  1). We thank the reviewer for these thoughtful comments and suggestions. Yes, it’s true that other shield designs exist (mostly in multiplate form); however, as we stated in the Introduction, they are expensive. There are also cheap versions of these multiplate (MP) radiation shields, as we have shown in Figure 1c on the manuscript; however, when we tested them, they did not perform well. Perhaps due to their smaller size, rainwater easily enters the sensor chamber, even during normal rainfall (without strong wind). Furthermore, these radiation shields are not particularly cheap, with marketplace prices reaching USD 28 or more. Based on these limitations, as the reviewer correctly guessed, we want to design a substantially cheaper radiation shield alternative that still works well for tropical urban environments.
  Per your suggestion, we revised the Introduction to clarify that although existing radiation shields, particularly commercial MP shields, can provide satisfactory performance, their cost may be prohibitive for dense urban monitoring networks. We also clarified that lower-cost alternatives often have durability limitations under tropical conditions, thereby motivating the development of a low-cost and robust alternative. The excerpt from the additional paragraph is shown below:
  “For distributed observation networks, MP radiation shields are commonly used because of their compact size and good measurement performance. However, high-quality commercial MP radiation shields remain relatively expensive when deployed in large numbers, whereas lower-cost alternatives often sacrifice durability and long-term weather resistance. Several low-cost radiation shield designs have been proposed, and many have demonstrated satisfactory measurement performance. However, prolonged exposure to intense solar radiation and wind-driven rain in tropical urban environments may limit the service life of inexpensive shields. Consequently, there remains a need for a radiation shield that combines low cost, robustness, and acceptable measurement accuracy for tropical urban deployments.
  To address the need for a radiation shield that is both affordable and sufficiently robust for dense urban monitoring networks, we designed and constructed …”.
  
  2). Thank you for your positive comments regarding the double-wall design. The double-wall design is not entirely our original idea. We followed the recommendations of the WMO for artificially ventilated shields (WMO, 2025a, p. 118). Originally, we designed the H-Tube radiation shield for much smaller and cheaper sensors, such as the Sensirion SHT4x and the Bosch BME280. However, to obtain a more accurate comparison, we decided to use the HMP155 sensor for the experimental tests.
  As the reviewer has noted, there is a difference in airflow speed inside the radiation shield when using smaller sensors (SHT4x) compared with the HMP155. We did not directly measure the aspiration rate inside the inner tube because we did not possess an anemometer or flow sensor that was small enough to fit inside the inner tube. As we described in the Discussion (Section 5.1), we measured the airflow at the inlet and outlet (fan) of the radiation shield and estimated the inner flow rate based on these measurements.
  For either the SHT4x or the HMP155, the outlet flow rate (where the fan is attached) was consistently around 3 to 3.3 m s⁻¹. However, the inlet flow rate differed depending on sensor size. For the HMP155, the inlet flow rate was around 1.6 to 1.8 m s⁻¹, while for the SHT4x it could reach 2.5 m s⁻¹. We speculate that the flow rate inside the inner tube could differ by up to 1 m s⁻¹ between the HMP155 and the SHT4x. Direct measurements of the inner-tube flow rate would, of course, be preferable to confirm this assumption.
  References:
  WMO: Guide to Instruments and Methods of Observation: Volume I – Measurement of Meteorological Variables, 2024 edition., World Meteorological Organization (WMO), Geneva, Switzerland, 601 pp., 2025a.
  
  3). Thank you very much for raising this question. This is perhaps one of the most important concerns regarding our design, and we spent considerable time evaluating it. To date, our radiation shield design has not been tested under actual strong tropical wind-driven rain conditions. While we have installed earlier prototypes outdoors, to our knowledge they have never been exposed to strong wind-driven rainfall, which typically occurs during the transition from the dry to the rainy season (usually beginning in November). This year's rainy season, however, may be delayed due to the strong El Niño event, according to BMKG (the Indonesian Meteorological Agency). We anticipate that we will be able to evaluate the effectiveness of the 45° inlet against actual wind-driven rain by January or February next year.
  In the meantime, we attempted to test the design using a hose and sprayer (see Figure R2a in the Response Figures). Water was sprayed almost horizontally from a distance of approximately 2 m. We estimated that the inclination angle of the water droplets from the vertical was approximately 75-80°, which could occur during rainfall with wind speeds exceeding 14 m s⁻¹ (Erpul et al., 2003). The radiation shield was sprayed for 5 min in both configurations, with and without the wire mesh. A nearby rain gauge indicated that the spray corresponded to a rainfall intensity of approximately 44-46 mm h⁻¹, which BMKG classifies as moderate rainfall.
  The results showed that, without the wire mesh, water could reach the inner tube (Figure R2a), suggesting that the design may not provide sufficient protection against nearly horizontal wind-driven rain. However, the addition of the wire mesh appeared to reduce the momentum of incoming water droplets, preventing them from reaching the inner tube and limiting water penetration to the inner surface of the elbow pipe (Figure R2b of the Response Figures).
  We acknowledge that these results are not conclusive, as a horizontal water spray directed continuously toward the inlet may not accurately represent actual wind-driven rainfall. During heavy rainfall, wind direction is often highly variable, particularly when associated with convective downdrafts. Under such conditions, our design, which has only two inlets, may have a lower probability of direct exposure than a multiplate (MP) shield, where airflow can enter from any direction. Conversely, it is also possible that the wire mesh may not always prevent the intrusion of wind-driven rain. The water droplets produced by the hose and sprayer were relatively large compared with the 3 mm mesh opening, whereas droplets transported by strong winds may be considerably smaller and thus more likely to penetrate the mesh. Therefore, only prolonged exposure to actual wind-driven rain events can adequately determine the effectiveness of the 45° elbow configuration. Should the 45° elbow prove insufficient, it could be readily replaced by a 90° elbow configuration similar to that proposed by Waugh (2021).
  Regarding the concern that the wire mesh may restrict airflow, we would like to clarify that all airflow measurements reported in our previous response were conducted with the wire mesh installed. Based on these measurements, we believe that the mesh does not substantially reduce airflow within the tube. However, as the reviewer correctly noted, water can accumulate on the mesh during wind-driven rainfall, as observed following the spray test (Figure R1b in the Response Figures). While this could potentially introduce a wet-bulb bias, we note that similar effects may also occur in conventional CRS (Stevenson screen) and MP radiation shields. Following rainfall, water droplets can often be observed on the louvres of a Stevenson screen and on the edges of the plates in MP shields. Therefore, susceptibility to moisture-related measurement bias may not be unique to our design.
  The reviewer's observation also highlights a broader measurement issue, namely the potential influence of radiation shield’s surface moisture on air temperature observations during and immediately following rainfall. The extent to which moisture retained on radiation shields affects measured temperature may warrant further investigation in future studies.
  References:
  Erpul, G., Darrell, N., and Gabriels, D.: The effect of wind on raindrop impact and rainsplash detachment, Transactions of the ASAE, 46, https://doi.org/10.13031/2013.12548, 2003.
  Waugh, S. M.: The “U-Tube”: An Improved Aspirated Temperature System for Mobile Meteorological Observations, Especially in Severe Weather, Journal of Atmospheric and Oceanic Technology, 38, 1477–1489, https://doi.org/10.1175/JTECH-D-21-0008.1, 2021.
  
  4). Thank you for raising this question regarding the wire mesh. The reviewer's concern about the wire mesh was also one of our concerns during the design of the radiation shield. Initially, we considered using stainless steel wire mesh, just like the material for the outer tube. However, we found that stainless steel wire mesh has different radiative characteristics from bright-annealed stainless steel pipe. Unlike the bright stainless steel pipe, stainless steel mesh may not reflect radiation as effectively. Therefore, we were concerned that it could act as a thermal mass and potentially introduce measurement errors.
  In addition, stainless steel mesh is relatively difficult to cut into a circular shape. Although laser cutting would facilitate fabrication, this approach proved to be relatively expensive. To avoid these problems, we replaced the stainless steel wire mesh with aluminium wire mesh, as shown in Figure R1b of the Response Figures.
  We selected the aluminium mesh because of its lower absorptivity, which we expect to minimize solar heating effects and reduce its influence on the measurements. We also inspected the mesh during periods of intense solar radiation and found that it remained relatively cool to the touch, suggesting that it is unlikely to significantly affect the measurement of air temperature and humidity. While we cannot completely rule out a small thermal influence of the mesh, its relatively small thermal mass, high ventilation rate, and distance from the sensor suggest that any such effect is likely to be limited.
  In the manuscript, we originally stated that stainless steel mesh was used. We have revised the manuscript to correctly indicate that the material was aluminium wire mesh. The revised sentence now reads as follows:
  “Additional protection was provided by end caps fitted with approximately 3 mm aluminium wire mesh (Figs. 2c, f, and Figs. 3d, h), which eliminated internal splash penetration during subsequent testing and also prevented entry of insects and debris. “.
  
  5). We are delighted that the reviewer asked about the size of the radiation shield. As we described previously, the current design was originally intended for small sensors such as the SHT4x. For such small sensors (<10 mm wide), we considered the dimensions of the shield (45 mm inner diameter and 77 mm outer diameter) to be more than sufficient to accommodate adequate airflow around the sensor, even during periods of rapid environmental change.
  We acknowledge that the larger dimensions of the HMP155 may influence airflow within the inner tube and potentially affect the sensor response during rapidly changing conditions. However, evaluating such effects was beyond the scope of the present study and would require dedicated testing under controlled conditions.
  We also considered developing a larger-diameter version of the radiation shield. One of our objectives is to accommodate multiple sensors, particularly air-quality sensors such as PM2.5 sensors. Our preliminary concept involves increasing the inner tube diameter to approximately 80 mm, allowing installation of sensors such as the Sensirion SPS series, which has a maximum dimension of approximately 41 mm. However, increasing the shield size would also increase material costs. Therefore, while we are considering a larger version of the shield, we will strive to keep the cost as low as possible, as affordability remains one of the primary goals of the design.
  
  6). Thank you very much for reminding us about the cable outlet of our H-Tube radiation shield. As depicted in Figure R3 of the Response Figures, in H-Tube_SS design, the cables for the sensors (either the HMP155 or SHT4x) and the aspiration fan exit through a hole located on the bottom side of the sensor mounting socket at the outlet side (see Figure 3c in the manuscript). For H-Tube_PVC, the cables exit through a similar hole on the bottom side of the outlet elbow pipe. All cables are located on the outlet side; therefore, they do not interfere with the airflow around the sensor, which is positioned toward the inlet side of the shield.
  Following the reviewer's suggestion, we revised the manuscript by adding a description of the cable outlet in the paragraph describing the mounting socket. The excerpt from the revised manuscript is shown below:
  “The inner tube and sensor were mounted concentrically within the outer tube using three-legged mounting sockets located at both ends of the assembly (Figs. 2c, f, and Fig. 3). The detailed geometry of the sockets is shown in Figs. 3a-c and Figs. 3e-g. The H-Tube_SS socket also functions as a connector between the outer tube and elbow pipe, whereas the H-Tube_PVC socket serves only as a support because the PVC elbow already incorporates a connector. One socket incorporates a sensor holder designed for the Vaisala HMP155A sensor; the design can be readily modified to accommodate other sensor geometries. This one socket also has a hole at its bottom side for the sensor cable to come out from the radiation shield. This design is intentional so that the sensor base and its cable does not affect the airflow to the sensor which located on the opposite side of the tube.”.
  
  Minor comments:
  Thank you for the suggestion to add references regarding the performance of low-cost multiplate (MP) radiation shields under wind-driven rain conditions. In these lines, we were describing the limitations of inexpensive marketplace MP shields, such as the example shown in Fig. 1c, rather than MP shields in general. We searched for relevant references but were unable to identify published studies that specifically evaluate the susceptibility of such low-cost MP shields to wind-driven rain. Our statement was based on field experience with several commercially available MP shield designs, ranging from low-cost marketplace products to higher-quality commercial models. We observed that many of the low-cost MP shields provided inadequate protection for the sensor during wind-driven rain events. In contrast, the better-performing designs typically incorporated an internal vertical tube within the shield chamber, effectively combining features of multiplate and vertical-tube shields. Given the lack of published literature on this specific issue, we have retained the statement but revised the manuscript to clarify that it reflects our field observations and practical experience rather than a conclusion directly supported by the literature. The excerpt of the revised paragraph is as follows:
  “… Although compact and portable, high-quality commercial non-aspirated MP shields often cost more than USD 200, while aspirated versions may exceed USD 700. Less expensive marketplace alternatives, such as shown in Fig. 1c, are available, but their durability and measurement performance are often poor. Based on our field experience, these low-cost designs perform poorly under challenging conditions, particularly strong solar radiation and wind-driven rain.”.
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2026-4598-AC1

Data sets

Junnaedhi_et_al_2026_H-Tube_RS Junnaedhi I Dewa Gede Agung, Suwarman, Rusmawan https://doi.org/10.6084/m9.figshare.33117401

Viewed

Total article views: 599 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
399	154	46	599	47	43

HTML: 399
PDF: 154
XML: 46
Total: 599
BibTeX: 47
EndNote: 43

Views and downloads (calculated since 24 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	203	40	18	261
Oct 2026	190	105	25	320
Nov 2026	6	9	3	18

Cumulative views and downloads (calculated since 24 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	203	40	18	261
Oct 2026	190	105	25	320
Nov 2026	6	9	3	18

Viewed (geographical distribution)

Total article views: 599 (including HTML, PDF, and XML) Thereof 599 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 04 Nov 2026

Short summary

Accurate urban temperature and humidity measurements are often limited by the high cost of radiation shields. This study developed an inexpensive H-Tube shield using common materials and 3D-printed parts. Field experiments showed that the aspirated design performed comparably to commercial products, while simple correction methods further improved accuracy. The H-Tube offers an affordable solution for distributed temperature and humidity monitoring networks in tropical cities.


Total:	0
HTML:	0
PDF:	0
XML:	0