Potential impact of iodine on tropospheric levels of ozone and other critical oxidants

A new analysis of tropospheric iodine chemistry suggests that under certain conditions this chemistry could have a significant impact on the rate of destruction of tropospheric ozone. In addition, it suggests that modest shifts could result in the critical radical ratio HO2/OH. This analysis is based on the first ever observations of CH3I in the middle and upper free troposphere as recorded during the NASA Pacific Exploratory Mission in the western Pacific. Improved evaluations of several critical gas kinetic and photochemical rate coefficients have also been used. Three iodine source scenarios were explored in arriving at the above conclusions. These include: (1) the assumption that the release of CH3I from the marine environment was the only iodine source with boundary layer levels reflecting a low-productivity source region, (2) same as scenario 1 but with an additional marine iodine source in the form of higher molecular weight iodocarbons, and (3) source scenario 2 but with the release of all iodocarbons occurring in a region of high biological productivity. Based on one-dimensional model simulations, these three source scenarios resulted in estimated Ix (Ix = I + IO + HI + HOI + 2I2O2 + INOx) yields for the upper troposphere of 0.5, 1.5, and 7 parts per trillion by volume (pptv), respectively. Of these, only at the 1.5 and 7 pptv level were meaningful enhancements in O3 destruction estimated. Total column O3 destruction for these cases averaged 6 and 30%, respectively. At present we believe the 1.5 pptv Ix source scenario to be more typical of the tropical marine environment; however, for specific regions of the Pacific (i.e., marine upwelling regions) and for specific seasons of the year, much higher levels might be experienced. Even so, significant uncertainties still remain in the proposed iodine chemistry. In particular, much uncertainty remains in the magnitude of the marine iodine source. In addition, several rate coefficients for gas phase processes need further investigating, as does the efficiency for removal of iodine due to aerosol scavenging processes.


the marine boundary layer (BL) [Singh et at., this issue]. (The possibility of interhalogen reactions serving as a sink for 03 is discussed below in the text.)
Investigations of bromine have also focused on the marine BL, but in this case the region of interest has been polar latitudes. Direct measurements of BrO radicals at levels as high as 17 pptv during Arctic springtime conditions [Hausmann and Platt, 1994] have indicated the presence of a significant BL source of reactive bromine. These measurements were also found to coincide with dramatic decreases in ambient ozone levels (e.g., from 40 to <0.5 parts per billion by volume (ppbv)) [Antauf et at., 1994]. These observations have led to considerable speculation about possible 03 destruction mechanisms, but in most cases these have been difficult to quantify due to the uncertainties associated with complex heterogeneous sources of bromine [Fintayson-Pitts et at., 1990;McConnett et at., 1992;Fan and Jacob, 1992]. Even so, the polar observations of BrO would seem to suggest that under some conditions, bromine could provide an important BL sink for 03.
The impact of iodine in the troposphere, like chlorine and bromine, has also received attention in the context of marine BL chemistry [Chameides and Davis, 1980;Cha•2etd and Crutzen, 1990;Jenkin et at., 1985;Jenkin, 1993]. More recently, it has been proposed as a sink for 03 in the lower stratosphere [Solomon et at., 1995]. Unlike bromine and chlorine, the production of reactive iodine requires no complex heterogeneous mechanism. For example, of the companion methyl halides, CH3I is the only species that readily photolyzes within the troposphere. In fact, its photochemical lifetime is roughly 4 to 5 times shorter than the most photosensitive bromine compound CHBr3.
The objective of this study will be to reexamine the role of iodine as a tropospheric sink for 03 and to also explore its ability to modulate the centrally important radical ratios HO2/OH and NO2/NO. This analysis has been motivated by both the availability of new CH3I data for the middle and upper free troposphere and the updated values for several critical photochemical/gas kinetic rate coefficients. The CH3I observations are the first of their kind and were recorded as part of NASA's western Pacific Exploratory Mission (PEM-West A).

Marine Observations of Iodine
As previously reviewed by Chameides and Davis [1980], there is now ample evidence showing that the methyl halide CH3I is a metabolic byproduct of many species of marine algae. This iodocarbon alone was estimated to provide a global source of iodine of the order of 1-2 Tg/yr. By comparison, industrial sources of volatile iodine were indicated to be less than 10% of the natural biogenic source. Yet other marine sources such as the direct photolysis of iodine compounds in ocean surface waters were listed as potentially important but still nonquantifiable.
More recently, Solomon et at. [1995] have provided an update of global marine iodine sources. The most significant additions have been those reported by Moore and Tokarczyk [1992], Ktick and Abrahamsson [1992], and Schatt and Heumann [1993].
Collectively, these investigators' observations indicate that the source strength for other iodocarbons (e.g., CH212, C2HsI , and CH2ICI ) could possibly exceed that for CH3I. For example, Schall and Heumann's measurements in the Arctic suggest that CH3I may only represent 26% of the total marine iodine released to the atmosphere. (Note that because of an expected red shift in the absorption spectrum for species such as CH2I 2 and C2HsI relative to CH3I, one would expect the lifetimes for these species to be even shorter than that for CH3I.) Atmospheric measurements of CH3I have primarily been confined to the marine BL with average mixing ratios ranging from 1 to 3 pptv [Liss and Slater, 1974;Singh et al., 1983;Rasmussen et al., 1982;Reifenhauser and Heumann, 1992]; but for regions characterized as having high biological productivity, much higher mixing ratios have been reported (i.e., 10-20 pptv), with one recent observation in the coastal waters of England reaching as high as 43 pptv [Oram and Penkett, 1994]. Marine areas nominally labeled as highproductivity regions are those typically found at high latitudes during springtime or those characterized as experiencing marine upwelling. In the latter regard, the equatorial upwelling in the Pacific could define an important area for the atmospheric influence of iodine. This conclusion reflects the fact that not only would this region be a potentially rich source of iodine but also would be a region that experiences considerable deep convection. In the first approach to verify the dilution procedure, a low pptv CH3Br standard, prepared by NIST, was compared against CH3I and CH3Br standards made from each of the pure compounds. In addition, permeation tubes for methyl iodide and methyl bromide were diluted in ultrapure air as a means of calibrating working reference gases. Good agreement was observed between the two methods. To estimate the precision of the CH3I measurements, 11 samples collected consecutively during a BL run and 7 consecutive samples collected around 7 km were analyzed. For the BL A to measure DMS included independent calibration runs for each individual measurement of DMS.) Since both compounds have predominantly marine sources, to a first approximation one might expect that the ratio of concentration for these two species should be similar at both high and low altitudes, provided their atmospheric lifetimes were not too dissimilar. In fact, at low altitudes some difference exists because DMS is controlled by OH oxidation and CH3I by photolysis (e.g., DMS <1 day versus CH3I 3-4 days); whereas at upper free tropospheric altitudes, reduced levels of OH and an enhanced UV flux result in nearly the same lifetime (i.e., -•2 days). For the PEM West A data set, a comparison of matched BL observations of DMS and CH3I (all above the limit of detection) gave a median ratio value for DMS/CH3I of 25. By comparison, for the highest-altitude data blocks, the value of this ratio was estimated at 43. Given the shift in lifetimes with altitude, one might actually expect that the high-altitude ratio should have been lower than that at low altitudes. However, upon reflecting on the fact that frequently there was a high degree of natural variability in the levels of both species and that the sampling integration times differed significantly (e.g., DMS, 3-4 min; CH3I, 10-60 s), one could argue that the level of agreement cited above still suggests that the high-altitude measurements of CH3I were probably not seriously compromised. Further evidence supporting this conclusion can be found in the level of correspondence between the DMS and CH3I measurements shown in Figure 1. This figure shows a scatterplot of these two species under conditions for which their lifetimes would be somewhat similar, i.e., data recorded for altitudes above 4 km. In this case a regression analysis, based on data taken from all latitudes, gave a Pearson R correlation coefficient of 0.71. Still the possibility of "offset" errors cannot be totally ruled out. At this time, therefore, our assessment is that the CH3I measurements are probably good to within a factor of 2to3. Figures 2a and  2b. These data have been binned every 2 km except for low altitudes where the height was confined to 1 km. In addition, both the upper and the lower 5% of the observations have been trimmed in the "range" display. The latitudinal zones shown are those designated elsewhere as "western tropical North Pacific," 0%18øN, and "western North Pacific rim," 18ø-42øN. This division reflects the PEM-West A O3 tendency analysis reported by Davis et al. [this issue].

The range and median values for the CH3I observations recorded during PEM-West A are shown in
The data displayed in Figures 2a and 2b show marine BL values of CH3I reaching as high as 1.5 to 1.8 pptv. For the tropics these levels are then observed to decrease to < 0.5 pptv. In contrast to the tropics, the Pacific rim data reveal the presence of many elevated CH3I observations at high altitude, with values at 10-12 km reaching >1 pptv. These high-altitude values reflect the influence of observations recorded during PEM-West A flight 9. This mission was unique in that the flight profile was designed to sample the low-altitude inflow and high-altitude outflow from Typhoon Mireille. The strong vertical pumping action from this typhoon produced enhanced high-altitude levels of both CH3 I as well as DMS [Newell et al.,  resulting hour glass shaped profile is thus seen to be similar to those for several other PEM-West A trace gases whose distributions were strongly influenced by deep convection [e.g., Blake et at., this issue]. By contrast, in the tropics, although BL values start lower (i.e., 0.6 pptv), a rather steady roll-off in CH3I median values occurs from the BL to altitudes of 8-10 km, resulting in a high-altitude median value of -0.15 pptv.

Vertical Transport and I x Distributions
As discussed below in the text with reference to the photochemistry of tropospheric iodine, the abundance of total reactive iodine Ix (I•= I +IO + HOI + HI + 21202 + INOx) in the middle and upper troposphere has been assessed in terms of a photochemical source, e.g., photolysis of CH31 and/or other iodocarbons, and losses in the form of surface deposition and heterogeneous uptake by cloud droplets, precipitation, and possibly aerosols. Thus given the stochastic nature of these loss processes in the middle and upper troposphere, one can estimate that the timescales involved could be significantly longer than that for vertical transport. We have investigated this question in this study by using a simple one-dimensional model [McKeen et. at., 1989] that incorporates a first-order diffusion process to describe vertical transport. A particular concern in employing the one-dimensional formulation is that it is difficult to envision how such a representation might adequately simulate deep convective processes that obviously are a major contributor to the relatively high mixing ratios of CH3I in the upper troposphere. In this regard we have assumed that the entrainment and detrainment of air by convective processes has been effective at nearly all altitudes. In other words, the combination of deep and shallow convection effectively mix boundary layer air to various altitudes. Moreover, we assume that the ensemble of air masses sampled during PEM-West A were representative of the ambient environment. As discussed in the above text, for the case of the tropics an examination of the median profile of CH3I (e.g., Figure 2a) as well as that of DMS does show a gradual decrease with altitude. Thus this is consistent with the above assumptions, and suggests that the median profile can be simulated with the eddy diffusion formulation. For this investigation a single value of the diffusion coefficient (K,) was used from 50 meters above the surface to the highest altitudes of the model at 12 km. Below 50 m a very stable layer was assumed with I x being immediately deposited at the surface upon contact. The initial concentrations of Ix and CH3I were assumed to be zero. An integration time of 20 days was chosen so as to permit CH3 I to reach a steady state condition at high-altitudes.
The key features of this one-dimensional transport simulation are illustrated in Figure Table 1, ("reference" curve), and (2) K z value of 80 m2/s with washout rates from Table 1 decreased by one half below 5 km and zero for altitudes above 5 km ("low washout").
a reasonable fit to the CH3I experimental observations was found when K z was assigned a value of 80 m2/s and the surface emission rate was given a value of 2x10 • molecules/m2/s. In this case, the value of Kz, relative to the photolysis loss rate, defines the vertical gradient for CH3I, while the flux value simply shifts the vertical profile along the abscissa. We note that our value for K z of 80 m2/s is only slightly larger than the value used by Liu et al. [1984] (i.e., 50-70 m2/s ). These authors based their value on observations of 222Rn over the continental United States during summertime. Like CH3I, the tracer 222Rn has only a surface source and during the summer season is probably also dynamically forced by convection to high-altitudes. The Ix "reference" profile shown in Figure 3 was obtained using the same K z value together with an altitude dependent washout rate for all soluble species as defined by Logan et al. [1981] (see Table 1). The resulting Ix profile shows a maximum of-0.5 pptv for altitudes approaching 12 km. To explore the magnitude of influence of the assumed value for the washout rate on Ix, a simulation was also carried out in which this rate was reduced by a factor of « for altitudes below 5 km and to zero above 5 km. As shown in Figure 3, Ix is enhanced by only 0.2 pptv, yielding a maximum value of 0.7 pptv. The effect resulting from assuming a much faster Ix removal rate than that employed for the "reference" profile is discussed below in the text.
For purposes of exploring the tropospheric impact of iodine in this investigation, we have selected three Ix source scenarios. The low scenario, involving 0.5 pptv of Ix at 10 km, assumes CH3I to be the only source of iodine and uses the "reference" value of K z of 80 m2/s and a CH3I flux of 2x10 • molecule/m2/s, e.g. low productivity marine water. A somewhat more realistic scenario, we believe, is one involving 1.5 pptv of Ix which reflects the inclusion of source compounds like CH2I 2 and C2H5I that could more than double the iodine flux to the atmosphere. The final scenario explored here involves a high-altitude value of Ix of 7 pptv. This scenario considers possibilities such as CH3I defining only 26% of the marine iodine flux to the troposphere (see earlier discussion of iodine sources), lower loss rates for Ix, and higher marine primary productivity relative to that observed during PEM-West A. We note that vertical profiles of Ix for scenarios 2 and 3 were obtained here by scaling the above cited 10-km values according to the Ix reference profile in

Description of Model
The basic photochemical model used in this study is similar to that described by Davis       iodine as related to the lower stratosphere. The chemical nature of these aerosols was taken to be predominantly sulfate and/or sulfuric acid, as are probably also aerosols for the upper troposphere. Although very little information is now available on heterogeneous reactions of iodine, Solomon and co-workers have argued from analogy with chlorine chemistry [e.g., Hanson et al., 1994] that because of the high reactivity of iodine, it would most likely not be permanently removed but would simply be recycled. Furthermore, these investigators suggest that the partitioning of iodine between the aerosol and gas phases would tend to favor most iodine being in the gas phase. However, given the impact of a 14-day aerosol sink, it would seem prudent to more fully explore the issue of heterogeneous loss of iodine.

Ozone Destruction
As discussed earlier in the text, the two most significant iodine-related 03 sinks involve the reaction sequences (R2),

(RS), (R14), and (R2), and (R9). Ozone destruction due to the IO self-reaction (R9) is found to be strongly dependent on both the altitude and the absolute Ix level selected.
Thus only for Ix levels at 7 pptv are IO levels found to be sufficiently high that this reaction contributes greater than 10% to the total iodine-related 03 destruction; for example, see Table 2. In addition, even at 7 pptv of Ix, (R9) does not become significant until reaching altitudes >6 km. For example, from Table 2 it can be seen that only at 10 to 12 km is approximately 30% of the iodine-related 03 destruction due to the IO self-reaction. Table 2 (top), for the case of 0.5 pptv Ix the predicted 03 destruction D(O3) due to iodine never exceeds that estimated from HOx alone by more than 8%. This leads to a relatively small 0 to 12-km column D(O3) effect of only 2%. Based on what we believe to be a more typical source scenario, involving 1.5 pptv Ix, Table 2 (middle) indicates that for some altitude/latitude ranges, as much as a 24% increase in D(O3) might occur. In all cases the general trend in D(O3) is seen as involving increasing iodine effects with A similar trend in D(O3) is seen for I• levels of 7 pptv, e.g., Table 2 (bottom). In this case, the upper altitude values of D(O3) are shifted by factors of 2 or more due to the inclusion of iodine chemistry. The resulting enhancement in total column 03 destruction is approximately 30%. If this evaluation of the PEM-West A tropical regime is further extended to the tropopause (based on generic high-altitude tropical trace gas compositions), we find that total 03 column destruction is increased by only an additional 2% over that cited above for 0-10 kin. This result once again makes the point that although very large effects can be found from iodine at high-altitudes, these altitudes tend to contribute less to the total column D(O3) effect.

As shown in
Taking for the moment the most extreme case for the impact of iodine on the troposphere (e.g., I• levels of 7 pptv), the total-column D(O3) results from Table 2 (Sx10 lø molecules/cm2/s) indicate that this photochemical destruction flux would be nearly 1.5 times larger than the 03 deposition flux to the ocean [e.g., Kawa and Pearson, 1989;Lenschow et al, 1982]. Equally noteworthy is the fact that, as discussed by Davis et al. [this issue], even before adding iodine, the PEM-West A tropical regime has been evaluated as having an imbalance between 03 photochemical production and destruction of between 3 and 12x10 lø molecules/cm2/s. (The lower deficit number assumes significant high-altitude (e.g., 10-17 km) production of 03 due the presence of high levels of lightning-generated NOx.) If, however, this region is at steady state for 03 as suggested by Davis et al., [this issue, 1995] then an additional 8x10 lø molecules/cm2/s 03 loss due to iodine raises still further questions concerning our level of understanding of the tropical ozone budget. Most likely, it means that during the time period of PEM-West A, I• levels were much closer to those labeled earlier as "typical," i.e., 1.5 pptv. Quite clearly, the optimum location for seeing a maximum iodine effect was not that defined by the PEM-West A mission as evidenced by the low primary productivity levels estimated from the low observed levels of CH3I. • I• levels at all altitudes have been scaled to the "reference" curve given in Figure 3. bColumn integrated numbers increasing altitude. However, it must be recognized that this trend is strongly influenced by the rapidly decreasing contribution to D(O3) from HOx chemistry alone. As a result, although the percent change in D(O3) is largest for high-altitudes, the total column effect on D(O3) is actually more influenced by middle free tropospheric altitudes.

Modulation of HO2/OH and NO2/NO Ratios
Based on (R8), (R14), and (R7), one would predict that with increasing levels of Ix, shifts should occur in both the ratios HO2/OH and the NO2/NO. But, as seen in Tables 3 (top) and 3 (middle) only minor changes are seen in these ratios when I x levels are limited to 0.5 and 1.5 pptv. By comparison, for I• levels of 7 pptv the HO2/OH ratio is observed to shift by as much as-18 to -24%, depending on latitude and altitude. The trend in this ratio is one which shows increasing negative values with increasing altitude. The average for all altitudes and latitudes is estimated here to be -15%. This leads to an average increase in the level of OH by 9%; for example, the OH effect typically defines 60% of the total change in the HO2/OH ratio. Given this shift in OH, if I• levels were to persist at the 7 pptv level for an extended period of time, one would also estimate a correspondingly 9% shorter lifetime for OH-controlled longlived trace gases such as CH4 and CH3CC13.
Concerning changes in the ratio NO2/NO, from Table 3 (bottom) it can be seen that 7 pptv of I• shifts the value of this ratio by a maximum of 11%. Furthermore, this shift is only observed for the highest altitudes examined. The average shift over all altitudes and latitudes is approximately 5%. Thus on average, the magnitude of the iodine shift in

Comparison With Previous Studies
The two most relevant tropospheric iodine modeling studies that have preceded this work have been those reported by Chameides and Davis [1980] and ChaOqeM and Crutzen [1990]. Chameides and Davis (C&D) only examined the case of the marine BL but did so for a wide range of median values of CH3I. Since only the total source of iodine is important, not the specific form, we have here been able to backout the level of CH3I for BL conditions that corresponds to our high-altitude 7 pptv Ix scenario. For this specific case, BL levels of CH3I would be in the range of 10 to 13 pptv. Based on this input, we find that the C&D standard model predicts an approximate 25% enhancement in D(O3) over that estimated from HOx related chemistry alone. As shown in Table 2 (bottom), this is basically the same enhancement predicted from this work (i.e., 20%) for BL conditions in the tropics. Although this agreement appears to be quite good, it turns out that it is actually quite fortuitous. For example, on closer inspection one finds that C&D estimated IO levels are nearly 1 order of magnitude higher than those estimated in this work. As a result, both the IO self-reaction and the reaction with HO2 are found to contribute significantly to D(O3). The reason for C&D's high IO levels now appears to be largely due to a major difference in the assignment of k values to the three critical processes (R6), (RS), and (R9). All three processes represent loss reactions for IO; and all three have now been assigned new k values based on more reliable kinetic and photochemical data. The new values are nearly 1 order of magnitude higher than the C&D values.
Cha•eld and Crutzen [1990] (C&C) also explored the question of tropospheric destruction of ozone by iodine although the main thrust of the study of these investigators was the possible impact of iodine on the oxidation of marine DMS. These authors concluded that iodine was probably an insignificant sink for ozone; however, this conclusion, was based on BL CH3I levels of only 3 pptv. In addition, C&C based their assessment on a k value for (RS) which has recently been reevaluated based on new kinetic data as being 1 « orders of magnitude higher [DeMore et al., 1994]. The J value for (R16) was also taken to be nearly 3 orders of magnitude lower than that estimated here. All 3 factors have been found to significantly minimize the tropospheric effects of iodine. This point was demonstrated here by carrying out simulations using PEM-West A median values for all photochemical species but invoking the C&C CH3I, (RS), and (R16) values. The D(O3) results in this case were within a factor of 1.5 of C&C, thus agreeing with the C&C evaluation in showing no significant effect on ozone from iodine chemistry.
As noted above, the availability of new kinetic data was the basis for the k value assigned to (RS) in this study. However, in the case of (R16) no spectroscopic data are available from which to evaluate this J value [Sander, 1986; denkin, 1993]. C&C assumed values ranging from 3x10 '4 to 2xl 0 '5 s '•, but this is 2 orders of magnitude lower than the J value calculated from the recently reported absorption spectrum for C1202 [DeMore et al., 1994]. We have taken what we believe is a rather conservative value for (R16), 9xJ(C1202) which is approximately 3 orders of magnitude greater than C&C.
This estimate is still 1 order of magnitude lower than the J value for (R6), involving the photolysis of IO.

Summary and Conclusions
The CH3I observations from PEM-West A, in conjunction with updated J and k values for several tropospheric iodine reactions, suggest that under certain conditions the addition of iodine chemistry to current models leads to predicted enhancements in the rate of destruction of ozone. It also predicts shifts in the critical free radical ratio HO2/OH. Although the percent effect on D(O3) was found to increase steadily with increasing altitude, the mid-troposphere was found to be the major contributor to the 0 to 12 km columnintegrated value of D(O3). The total column effect on D(O3) was estimated to be only 6% for Ix levels of 1.5 pptv; however, this increased to 30% at tropical latitudes for the high Ix scenario involving 7 pptv. A potential enhancement in D(O3) of 30% represents a major impact on the 03 budget at tropical latitudes. In fact, it is difficult to reconcile with the findings of Davis et al. [this issue] who report that for the time period of PEM-West A this region displayed significant net-photochemical column destruction even in the absence of iodine chemistry. This most likely suggests that during PEM-West A, Ix never reached the 7 pptv level. This conclusion would seem to be in agreement with the observations that BL levels of CH3I were also low indicating a marine region having low primary productivity.
The influence of iodine on the radical ratio HO2/OH was found to be most significant for the upper free troposphere and for I,, levels of 7 pptv. For this I,, level the average effect for all altitudes was approximately 15%. The effect of iodine on the NO2/NO ratio was also found to maximize at high-altitude but never exceeded 11% even for the highest Ix scenario. If the maximum estimated shift in the HO2/OH ratio were to be realized, this would result in a 9% increase in atmospheric levels of OH and thus lead to an approximate 9% decrease in the lifetimes of OH-controlled trace gases such as CH 4 and CH3CC13.
Of interest also is the fact that median levels of CH3I in the 8-10 km altitude range (e.g., for the western tropical Pacific) reached 0.15 pptv, with individual observations reaching 0.3 to 0.7 pptv. Although these altitudes are still significantly removed from the tropical tropopause, if the mixing ratios observed for CH3I represent only a fraction of the total iodine being convectively pumped into the upper troposphere, it is still possible that Solomon et al. 's [1995] requirement of a few tenths of one pptv of iodine in the lower stratosphere might be reached on some occasions. In this context, because of the high sensitivity for detection of iodine species like CH3I, together with its short lifetime of 2-4 days, CH3I would appear to be an excellent tracer for the transport of marine BL air into the upper troposphere and lower stratosphere.
Major uncertainties in the proposed iodine chemistry include the total marine iodine source strength as well as the reliability of several k and J values. In particular, it will be important in future field studies to include not only measurements of CH3I but also those for other iodocarbons. In addition, both the critical IO reaction involving HO2 (RS) and the self-reaction (R9) should be further investigated, as should the photochemical cross section for the species I202. Finally, the efficiency for removal of iodine species due to high-altitude aerosol must be resolved. Quite significant to gaining more direct insight into all aspects of the proposed iodine chemistry would be direct observations of the radical species I and IO.