the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Jun 2026
SPS30 and SEN55 PM2.5 sensor intercomparison and validation through indoor and outdoor measurements in Arba Minch, Ethiopia
Abstract. Ethiopian air pollution is understudied yet highly relevant considering population size and source abundance including solid fuel cooking, small scale waste burning and vehicle fleet. Low cost PM2.5 sensors can be used to mitigate this by mapping pollution exposure. We report on indoor and outdoor measurements with Sensirion sensors SPS30 and SEN55, the former extensively validated in literature, the latter not. We evaluate their use in Arba Minch, population ca 200.000. In addition to sensor inter and intra comparison we benchmark the low-cost sensors against gravimetry. Furthermore, a separate Swedish outdoor dataset is included to extend the range of particle types and loadings and evaluate relative humidity (RH) effects using a reference-equivalent monitor (Palas FIDAS).
We found that the SEN55 consistently reports values 6–10 % higher than the SPS30; once this systematic offset is corrected, the sensor types are functionally identical with high precision (coefficient of variation ≤ 7.7 %, between-sampler uncertainty ≤ 1.7 µg m-3). Both sensors demonstrated high stability across repeated high-concentration events (> 1000 µg m-3). While the SEN55 exhibits digital truncation at 6553.4 µg m-3, we find that measurements beyond the 1000 µg m-3 manufacturer specification remain meaningful and essential for accurate mass estimation in biomass-burning environments. The impact of relative humidity was small and consistent across both sensor types.
Our study shows that the SPS30 and SEN55 – when calibrated under circumstances of use – are stable and accurate instruments (indoor and outdoor accuracy error ≤ 22 %, outdoor expanded uncertainty ≤ 15 % in comparison to gravimetric measurements). Pragmatic, large-scale low-cost monitoring supported by mobile gravimetric validation offers the most viable path toward mitigating air pollution exposure in resource-constrained settings.
- Preprint
(4661 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 16 Aug 2026)
-
RC1: 'Comment on egusphere-2026-2537', Anonymous Referee #1, 13 Jul 2026
reply
-
CC1: 'First reply on RC1 _ Q5', Johannes Dirk Dingemanse, 17 Jul 2026
reply
To facilitate a productive discussion, I (corresponding author) reply already on parts of the RC1 questions. I however do so as community comment, to distinguish this from a later full response on behalf of all authors. This comment is regarding RC1 fifth question: "Appendix A.3: The post-hoc exclusion of SEN1 and SEN2 in Phase 2 is justified only by citation, not by the authors' own data. Please report the affected results with and without these exclusions, and state whether the 12.5% SDT rejection threshold was defined a priori."
In the attached PDF I included a figure (Fig.1) and a table (Table 1) that includes SEN1 and SEN2 data.
SEN1 shows an overall wide variation (Fig. 1). This leads to an r of 0.53, and an RMSE significantly higher than any of the other sensors (Table 1). SEN1, as part of Box1, was the only sensor positioned perpendicular rather than front-facing, and it has seen different orientations versus source and collocated UPAS across the measurement period (the second half, more sensors were added to the same location, so the box was repositioned). We expect this is the reason for this wide variability. After bias correction, the AE is higher than 25% (44%). SEN2 simply has little data (n=6). It is quite linear (r=0.91), and it is following the SEN3 (both were in the same Box2) data quite well (slopes of 1.12 and 1.11, respectively). However, the variation versus number of data points is simply too high to achieve an AE < 25% (39%).
When including the SEN1 and SEN2 data into the metrics for all sensors combined, the linearity is weaker (r 0.87 instead of original 0.93; R2 0.95 instead of 0.97, RMSE 264 instead of 187), but the ultimate slope is similar (1.15) and AE is still < 25% (20%). In other words: SEN1 and SEN2 results add to the random error, but do not alter the conclusions. Given that the SEN1 and SEN2 errors are connected to reasons other than the sensor quality (SEN1: other orientation, SEN2: significant dataloss due to power supply problems), we prefer to leave it out of the main text, as it does not speak to the sensor quality but just to field challenges and choices.
-
CC2: 'First reply on RC1 _ Q3', Johannes Dirk Dingemanse, 17 Jul 2026
reply
To facilitate a productive discussion, I (corresponding author) reply already on parts of the RC1 questions. I however do so as community comment, to distinguish this from a later full response on behalf of all authors. This comment is regarding RC1 third question: "Section 2.1 and Appendix D: Outdoor 2 is located behind a window (physically indoors) and contributes 36 of the 53 outdoor filter pairs. Indoor placement attenuates concentration peaks and may modify the sampled size distribution, and Appendix D.1 already shows different variability at the two sites. Please report the regression results separately for Outdoor 1 and Outdoor 2 to justify pooling them, and describe the nature of this site in the main text rather than only in the appendix."
In the attached PDF I included a figure (Fig.1) and two tables (Table 1, 2) that show data for locations Outdoor 1 and 2 separately.
The data points for Outdoor 1 and 2 are visibly in the same range and of similar trend (Fig. 1). For each sensor, the confidence intervals (CIs) of Outdoor 1 slopes overlap with those of Outdoor 2 (Table 1), and of both locations they overlap with the CIs for pooled data (manuscript Table 5). Furthermore, the difference between SPS30 and SEN55 runs across both locations (with SEN55 having a lower slope i.e. less of an underestimation). If there is a difference to speak of, it is that at Outdoor 1 slopes are lower than at Outdoor 2. But, also, that the Outdoor 1 data is slightly more variable: accuracy error (AE) and expanded uncertainty (WCM) after correction is higher than the original situation – while those of Outdoor 2 are lower. The higher Outdoor 1 variation also results in the SEN55 versus SPS30 slopes not being significant (CIs overlap).
When applying the slopes based on the pooled location data as correction factors to the sensor data (1.23, 1.21, 1.20 and 1.12 for SPS1, SPS2, SPS3 and SEN1, respectively), and then calculating AE and WCM for each sensor, the maximum AE and WCM for Outdoor 1 are 16% and 22%, and for Outdoor 2 10% and 11% (Table 2): still lower than the 25 % thresholds. Even when applying the averaged SPS30 slope (1.21) as correction to all individual sensors, all resulting AEs (including that of the SEN55) are < 25 %. The WCM for the SEN55 becomes 27 %, which still is well below the threshold for indicative sensors (< 50 %) – but not below that for reference-grade (< 25 %).
In other words, while there might be a difference in sensor to gravimetry for Outdoor 1 and Outdoor 2, we do not have conclusive evidence (non-significant slope differences), using results of the pooled data does not lead to unacceptable accuracy errors for either of the locations, and the main conclusions still stand: the sensors can reach required accuracy after site-specific calibration, SPS30 and SEN55 are similar enough to allow for extrapolating data quality results of the SPS30 to the SEN55, but a better accuracy is reached when taking into account differences between SPS30 and SEN55.
-
CC1: 'First reply on RC1 _ Q5', Johannes Dirk Dingemanse, 17 Jul 2026
reply
Interactive computing environment
Data and code for SPS30 SEN55 PM2.5 indoor outdoor validation, Arba Minch, Ethiopia Johannes Dirk Dingemanse https://doi.org/10.17605/OSF.IO/EC3T8
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 134 | 46 | 12 | 192 | 7 | 6 |
- HTML: 134
- PDF: 46
- XML: 12
- Total: 192
- BibTeX: 7
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please see the attached file for detailed comments.