Along with information related to the rearing of approximately 1.5 million IIT or IIT/SIT derived male mosquitoes [10], approximations of Hawaiʻi specific infrastructure, utility, and personnel costs were used in this analysis. The calculator presented here uses this compiled information to assess infrastructure costs associated with developing a C. quinquefasciatus rearing facility at a capacity needed to control known densities in an area the approximate size of the Alakaʻi Wilderness Reserve. The Alakaʻi Wilderness Reserve is within the Na Pali-Kona Forest Reserve, a reserve that includes known forest bird refugia, such as Kauaʻi’s Alakaʻi swamp (Fig. 1). We assume here that a 2km2 buffer around the wilderness reserve (as derived from [22]) will minimize emigration of C. quinquefasciatus from outside source populations. This 2km2 buffer represents the maximum mean distance traveled by C. quinquefasciatus [23]. Table 1 lists the variables used in the code to itemize the infrastructure costs and a description of each variable. Variables are partitioned into three main types in Table 1: those variables associated with ecology of Hawaiʻi and the Culicid (Mosquito and Site Information), those itemizing the basic infrastructure costs (Basic Infrastructure Costs), and those associated with position and personnel expenditures (Position and Personnel Costs). We developed this analysis in the R statistical environment [24] (Code S1, Supplemental Materials).
Table 1
Variable names and descriptions used in the R code to derive an estimate of infrastructure and personnel costs.
CODED VARIABLE | VARIABLE DESCRIPTION |
MOSQUITO AND SITE INFORMATION |
MaunaLoa.Mos.Sites | Sites from Samuels et al. (2011) in which estimates of density were defined |
MaunaLoa.Mos.Density.km | Mosquito Densities (individuals/sqr. km.) at the various sites from Samuels et al (2011). |
MaunaLoa.Elev.m | Approximate elevation of sites outlined in the *MaunaLoa.Mos.Sites* input in meters |
Alakai.Area.ToCntrl.km | Approx. area of the Alakai Wilderness Reserve with a 2 kilometer buffer, in square kilometers |
FemalePercent | The female percent of the sex ratio. A value of 50 indicates an equal sex ratio. |
Overflooding_Multiplier | Multiplier to derive the overflooding ratio needed for an effective control strategy. A value of 10 indicates that to suppress the wild-type population a successful program needs 10x the number of SIT/IIT laboratory derived Culicid. |
BASIC INFRASTRUCTURE COSTS |
Year1.Only.Items | Infrastructure items needed in year 1 for rearing approx. 1.5 million IIT/SIT male Culicid, not including personnel |
Year1.Only.Costs | Approx. costs of *Year1.Only.Items* for rearing 1.5 million IIT/SIT male Culicid |
Electricity | Yearly electricity costs to rear approx. 1.5 million IIT/SIT males |
LaboratorySpace | Cost of a 800 meter squared laboratory space as needed for rearing approx. 1.5 million IIT/SIT males |
AllOtherYear.Items | Misc PCR/Lab/Field Supplies for items used in each year of the control application (e.g blood) |
AllOtherYear.Costs | Approx. costs of *AllOtherYear.Items* for items used in each year of the control application |
POSITION AND PERSONNEL COSTS |
Personnel.Des | Types of technician positions to be funded |
Wage.Mass.Rearing | Average hourly wage for the Mass Rearing Technician position |
Wage.Quality.Control | Average hourly wage for the Quality Control position |
HoursPerYear | Yearly hours for each position |
Fringe | Percent Fringe costs for full time employees |
Biological/Ecological Characteristics of C. quinquefasciatus used for the analysis
As noted in Table 1, mosquito densities were derived from [5] (see Appendix B Table 1B in [5]) and defined in that paper as the number of individuals per km2 for each site. Infrastructure costs were determined for all densities of C. quinquefasciatus at each elevation in which they were present as it is assumed that C. quinquefasciatus densities vary based on temperature [5], and elevation can be used as an imperfect proxy for the variance in temperature between sites. Conducting the analysis such that it takes into account this variance in elevation allows the user of the information to better estimate site and cost variance, as well as the effect of sex ratio.
the male to female sex ratio was maintained at 1:1 (equal sex ratio) for most analyses conducted in this assessment. As a test of how robust the assessment is to variation in the sex ratio parameter, we compared an analysis with the default equal sex ratio (equal) to an assessment conducted with a female biased sex ratio.. Sex ratios commonly seen in the literature for C. quinquefasciatus (or other Culicid) vary [25–27], but are within the realm of both ratios used. While maintaining all infrastructure, personnel and density costs, a comparison between an equal (1:1) and female biased sex ratio (70% female, or 0.43:1 male:female ratio, the converse of that from [27]) was conducted to illustrate the change in cost associated with this variable. Operations are optimized at a mass rearing facility such that sex ratios are equal.
Infrastructure Costs
As noted in the methods overview most first year, and subsequent year, infrastructure costs were estimated using information published (as an average cost/km2) in [10]. One large purchase that was not itemized by these estimates was the Arthropod Containment Level 2 (ACL-2) rearing space [28] necessary to rear approx. 1.5 million male Culicid. It was determined that at maximum a facility must be ~ 300-500m2 (enough to produce 500k − 1 million male SIT/IIT Culicid) to be cost effective, and allow for future scaling efforts. Therefore an 800m2 ACL-2 facility (Table 2: LaboratorySpace) would be of sufficient size to rear approx. 1.5 million male C. quinquefasciatus. However, the cost could vary from this estimate depending on the company used, the facility type, and previous ownership (versus new) etc. The cost of an 800m2 facility was estimated using three converted (from ft2 to m2) median cost quotes (Quotes S1, Supplemental Materials), as defined per m2, and multiplied by the minimum size of a facility (800m2).
Table 2
Variable names and values used in the R code to derive an estimate of infrastructure and personnel costs.
CODED VARIABLE | DEFAULT VARIABLE VALUE |
MOSQUITO AND SITE INFORMATION |
MaunaLoa.Mos.Sites | Malama Ki; Nanawale; Bryson's; Waiakea; Cooper; Crater; Pu'u |
MaunaLoa.Mos.Density.km | 4,546; 78,547; 14,597; 29,001; 27,615; 1,637; 618 |
MaunaLoa.Elev.m | 25; 36; 314; 885; 1,024; 1,177; 1,247 |
Alakai.Area.ToCntrl.km | 117 kilometers squared |
FemalePercent | 50% (Assumes an equal sex ratio) |
Overflooding_Multiplier | 10 (10:1 overflooding ratios are commonly used for a control efficacy of 99% (Zheng et al. 2019; Kandul et al. 2019)) |
BASIC INFRASTRUCTURE COSTS |
Year1.Only.Items | Irradiator; Mosquito.Sex.Sorters (6); larvae Rearing Units (5); adult cages (100); ovitraps (300); BG traps (50); PCR Machine |
Year1.Only.Costs | $200,000; $6,900; $134,500; $11,040; $2,400; $7,500; $47,000 |
Electricity | $2,000.00 x 12 |
LaboratorySpace | median price of a 800 meter squared modular facility (see Supp. Mat. Section 2 for price quotes) |
AllOtherYear.Items | PCR Buffers, reagents, primers, Taq, Misc. Equipment, Misc. Field Supplies. |
AllOtherYear.Costs | $30,000 |
POSITION AND PERSONNEL COSTS |
Personnel.Des | Mass Rearing; Quality Control |
Wage.Mass.Rearing | $20.00 |
Wage.Quality.Control | $25.00 |
HoursPerYear | 260 * 8 |
Fringe | Research Corporation of the University of Hawaiʻi Fringe/Indirect is set at 61.56% for 2018. |
Another high but potentially optional item in the year 1 cost is the irradiator used in IIT/SIT to make the females infertile at such low doses as to not affect male fitness [10]. These costs may vary if the release program is able to use a machine learning/artificial intelligence adult sex selection discriminator, such as that developed by Verily Life Sciences [29]. It is important to note that at this point the cost of the irradiator may be lower than the cost of the Verily technology, but those costs may change with economies of scale. To illustrate the cost variance of the irradiator (Tables 1 & 2) the calculator was run with and without the irradiator costs. The highest perpetual costs (year 1 and beyond) are those associated with rearing and quality control personnel (Tables 1 & 2). To rear approx. 1.5 million adult male Culicid every week requires eight rearing and three quality control technicians for the extent of the work year (260 days, 8 hours/day, Table 2). In the calculator presented here, the default cost per hour of these different positions was higher for the quality control position as it is primarily managerial (Table 2). The wages used here are likely on the low end of the wage spectrum and should be modified as appropriate.
Ratio Calculation
In this calculator, the larval rearing units are treated as the primary delimiter defining the production scale of a Culicid mass rearing facility. In other words, each incremental increase in the number of larval rearing units necessitates a certain amount of space (for laboratory work, adult rearing, sex sorting, office space, utilities, etc.), positions, and other costs outlined in Table 1. We also assume that each larval rearing unit can rear ~ 1 million Culicid/week. Under this assumption, to house 1 rearing unit or produce ~ 1 million Culicid/week (Table 1) [21, 30],and the associated staff etc., a facility would need at maximum 300-500m2 of laboratory space; to house 5 rearing units and the associated staff etc., it is assumed that a facility would need at maximum a 800m2 laboratory space. For each projection, the rate of increase is calculated and supplied. These rates can be thought of as both the number of larval rearing units needed as well as the rate of increase for all other items (space, positions etc.) associated with rearing the necessary density of male Culicid (Table 2). Rates are rounded up to the nearest whole unit, from two significant digits of the proportion of wild-type males to laboratory males needed. Conducting the assessment in this way ensures the production capacity needed for the successful implementation of an IIT or IIT/SIT control program. The simplistic but essential equation ((1)) used to define the rate linking production capacity to cost is below:
$${r}_{t}=\frac{wmo}{{u}_{t}{n}_{t}} \left(1\right)$$
In Eq. (1) the rate of increase at a specific time point (\({r}_{t}\)) is defined by the interaction of the wild type Culicid population size (\(w\)), the proportion of males in the wild Culicid population targeted (\(m\)), and the anticipated overflooding ratio (\(o\)). These (\(wmo\)) are then divided by the capacity of a single larval rearing unit as defined for a specific time point (\({u}_{t}\)) multiplied by the proportion of laboratory males reared by that rearing unit at that time (\({n}_{t}\)). For all projections estimates were developed using the maximum weekly production (\(t\)) of IIT males per rearing unit [21, 30].
Additional Examples Of Utility And Reproducibility
In order to maximize the utility of the calculator, and enhance its understanding, two additional examples were completed. The first looks at C. quinquefasciatus infrastructure needs to develop IIT or IIT/SIT in the home range of ʻiʻiwi D. coccinea (Example report S1, Supplemental Materials), a native Hawaiian Honeycreeper federally listed as threatened [31]. The second example shows how, with just a few slight modifications, the calculator can be used to assess the efficacy of IIT or IIT/SIT for control of additional Culicid. In this example, the life history characteristics of A. aegypti are used instead of C. quinquefasciatus to assess infrastructure needs for IIT or IIT/SIT control on the island of Hawaiʻi (Example report S2, Supplemental Materials). For both examples, we provide a sample report and all modified R code. Additionally, we wrote this manuscript and both example reports in Rmarkdown to maximize reproducibility [32–34] (Code S2, Supplemental Materials).