This manuscript reports an interesting investigation of radical budgets (OH, HO₂, RO₂ and RO_x) in ambient air flushed through the SAPHIR chamber. This chamber is well equipped to measure each radical - or group of radicals - together with ancillary data that is necessary to compute the radicals' production and destruction rates.

The authors report on measurements covering each season of 2019 and focus on the spring and summer seasons when radical concentrations were significantly larger than instruments’ detection limits. Great care was taken to rule out potential instrumental issues on the measured radical concentrations. Contrasting production and destruction rates for the above-mentioned radicals, the authors highlighted significant gaps in our understanding of radicals’ sources and sinks in air masses characteristic of the Julich atmosphere (urban area impacted by biogenic emissions). It is shown that at low and moderate levels of NOx, an additional source of ROx is needed to close the RO_x production-destruction balance. At low NO (< 1ppb), it is shown that additional processes leading to the production of OH and the destruction of HO₂ are needed to close the budget of theses radicals. At higher NO (1-3 ppb), no imbalance is observed for OH and HO₂ but an unknow source of RO₂ is needed to balance their production and destruction rates. The authors did a great job assessing how uncertain chemical processes could contribute to the observed imbalances.

This reviewer thanks the authors for the effort put in providing an easy-to-read manuscript - very well structured, clear and concise information - which is not an easy task when dealing with such complex dataset of measurements.  This work will be of great interest for the atmospheric community and this reviewer recommends publication after the authors address the following comments:

Main comments:

P8 L204-207:  It is stated that the ROx-LIF system is calibrated for CH₃O₂. Since the sensitivity of this instrument is species dependent, this will lead to a measurement bias for RO₂. By how much could the measured RO₂ deviate from the true value? Is this bias factored in the measurement uncertainty? If not, how could it affect the calculations of P_OH,isop, (Eq. 4), P_HO2 (Eq. 6), D_HO2 (Eq. 7), D_RO2 (Eq. 11) and D_ROx (Eq. 13)?

P10 L306: “First, the contributions from CO, NO, NO2, HCHO and O3 is removed from the measured OH reactivity as these species do not form RO2 radicals in the reaction with OH. It is then assumed that the remaining fraction can be attributed to organic compounds (VOC reactivity (kVOC)) including measured and unmeasured VOCs, which produce RO2 radicals in their reaction with OH” – For some VOCs the reaction with OH can lead to the prompt formation of HO₂ together with RO₂. For instance, toluene+OH will form 28% HO₂ and 72% RO₂. Assuming only the formation of RO₂ could lead to an underestimation of P_HO2 and an overestimation of P_RO2. Could the authors comment on this aspect? Can the prompt formation of HO₂, which occurs with a few VOCs, be neglected when the total pool of VOCs is considered?

Table 2: This table indicates that NO₂ was measured using a chemiluminescence instrument. Was this instrument equipped with a photolytic NO₂ converter or a molybdenum converter? This should be clearly stated. Instruments equipped with a molybdenum converter are known to be prone to interferences when measuring NO₂. If a molybdenum converter was used, the authors should discuss how interferences on NO₂ measurements could impact the calculations of ROx destruction rates (Eq. 13) and Ox production rates (Eq. 14).

Minor comments:

P8 L226-228: “Photolysis frequencies inside the chamber were derived from the solar actinic flux densities measured by a spectroradiometer mounted on the roof of the nearby institute building (Bohn et al., 2005; Bohn and Zilken, 2005).” – How is the Teflon sheet transmission determined when the cleanliness changes from day-to-day?

P11 L338-339: Was the humidity dependence of k17 accounted for in the calculation of D_ROx? It seems so from Table 1 where k17 is reported for 1% water but it should be clearly stated in the text.

P12 L370-371: Please include O3+alkenes in the list of minor Ox destruction pathways

Figures 3, 4, S3, S4: Please indicate in the caption what the error bars represent for OH. Also please add error bars for the other measurements.

Figures 8-12 & S6-S7: Please clarify in the caption whether uncertainties are displayed as 1-sigma? 2 sigma? Other?

Edits:

P2 L55: “the lower are summarized in” should read “the lower atmosphere are summarized in”

Table 1: The authors may want to add the reaction of OH+O3 in the radical interconversion section

P32 L819: “the formation of OH from their reaction with NO could only explain up to” should read “the formation of OH from their reaction with HO2 could only explain up to”

This paper presents measurements of OH, HO2, and RO2 radical concentrations and total OH reactivity in ambient air sampled from inside the SAPHIR chamber. OH radicals were measured by both an LIF-FAGE instrument and a DOAS instrument, while HO2 and RO2 radicals were measured using the ROx-LIF instrumental technique. The authors use the radical measurements together with measurements of ancillary measurements of photolysis rates, NOx, and VOCs to calculate the radical production and destruction rates to determine whether they are balanced. The authors find that under lower NO conditions (less than 1ppb) the rate of OH radical destruction was consistently greater than OH production, while the rate of HO2 radical production was greater than HO2 radical destruction. At higher NO conditions, the OH and HO2 radical budgets appeared to be in balance. In contrast, the total ROx radical budget was found to be generally balanced over the range of NO measured, with the destruction rate greater than the production rate at the highest and lowest mixing ratios of NO.

The authors demonstrate that instrumental errors associated with the radical measurements are minimal given that the LIF-FAGE measurements of OH (after accounting for interferences) were in good agreement with the DOAS measurements, and that the measurements of the rate of Ox production was consistent with the rate of the reaction of peroxy radicals with NO. As a result, the authors conclude that a missing OH radical source and a missing HO2 loss process is needed to balance the radical budgets. At the highest NO concentrations, the authors suggest that an additional RO2 production process is required to close the RO2 radical budget.

The authors suggest several possible explanations for the missing sources and sink, including alkene ozonolysis, OVOC photolysis, and heterogeneous uptake of HO2 radicals, although according to the conclusions the exact nature of the missing sources and sink could not be determined from the measurements.

The paper is well written and contains additional evidence that our understanding of radical chemistry under a range of NO concentrations is incomplete. The paper would be suitable for publication after the authors have addressed the following comments.

• In the abstract (lines 39-41), the authors state that the missing OH source “consists likely of a combination of a missing primary radical source (0.5 ~ 1.4 ppbv h^-1) and a missing inter-radical HO2 to OH conversion reaction with a rate of up to 2.5 ppbv h^-1.” However, there appears to be little discussion of this potential OH source/HO2 sink in the paper, except briefly on page 29 (lines 670-671) and page 43 (line 893) and it is not mentioned in the conclusions. If this is a major finding as suggested in the abstract, it should be emphasized more in the manuscript.
  
  
  
  
• The authors state that photolysis frequencies were “were derived from the solar actinic flux densities measured by a spectroradiometer mounted on the roof of the nearby institute building.” Given that an underestimation of radical production from photolysis could account for the missing OH radical source, the authors should clarify how potential differences in photolysis rates inside versus outside of the chamber were accounted for in their budget calculations.
  
  
  
  
• The authors measured total OH reactivity and use it to determine the total OH loss rate. However, as illustrated in Figure 5 there appears to be significant missing OH reactivity when compared to the calculated reactivity from measured OH sinks. Unfortunately, there is little discussion about the potential composition of the missing OH reactivity. The paper would benefit from a brief discussion of the missing OH reactivity and whether unmeasured OVOCs may be responsible. While the authors suggest that OVOCs such as acetaldehyde, methyl vinyl ketone, methacrolein, and methylglyoxal do not contribute significantly to radical production, have the authors considered other potential unmeasured OVOCs, perhaps through a model of the chemistry, that may be contributing to the missing reactivity as well as be a potential unmeasured radical source?

Minor comments:

Page 2, line 63: The Griffith et al. 2016 reference reports urban measurements. Did the authors mean to cite Griffith et al., Atmos. Chem. Phys., 13, 5403–5423, 2013, which reports measurements in a forest environment?

In Figure 12 I assume that the numbers at the top of the figure represent the number of points in each NO bin. This should be clarified in the caption. Also, the uncertainty should be clarified.