1Purpose and scope
KiwiPest helps a grower, forester, farm adviser, restoration practitioner or student choose among the pesticides used in New Zealand for a given problem, by comparing them on the properties that decide their environmental and human risk at a particular site. It covers herbicides, insecticides, acaricides, fungicides, bactericides, nematicides and soil fumigants, quarantine fumigants, molluscicides, vertebrate toxic agents and plant growth regulators, across pastoral, arable, vegetable, orchard, kiwifruit, viticulture, forestry, nursery, turf, non-crop, biosecurity, aquatic and native-restoration land uses.
It is a screening tool in the tradition of the pesticide indices used by regulators and extension services (the PPDB hazard classes, the Groundwater Ubiquity Score, the FOCUS drift tables, the EPPO bee hazard quotient). It ranks risk, not efficacy: an active that ranks well may still be the wrong choice if it does not control your pest at your crop stage, and it says nothing about which product label covers your situation.
The label, the EPA approval conditions, the NZ Food Notice: Maximum Residue Levels and your regional plan govern every use. KiwiPest's withholding periods are indicative and must never be used in place of the label.
2The database
Each active ingredient carries the properties below. Values are typical (often the regulatory "agreed endpoint") literature figures, taken mainly from the University of Hertfordshire Pesticide Properties Database (PPDB), EU and US EPA assessment reports and the NZ EPA's own assessments; see Data sources & terms.
| Property | Unit | Meaning |
|---|---|---|
| Water solubility | mg/L at 20 °C | How much dissolves; high solubility favours movement in water. |
| Koc | mL/g (= L/kg) | Sorption to soil organic carbon. Converted to a site Kd in §5. |
| Kd (inorganic actives) | L/kg | Copper has no Koc; a typical soil Kd is tabulated instead. |
| Soil DT50 | days at ~20 °C | Time for half the active to degrade in aerobic soil. |
| Water DT50 | days | Water or water–sediment system half-life, where reported. |
| log Kow | — | Octanol–water partition: above about 3–4, a bioaccumulation concern. |
| Rat oral LD50 | mg/kg body weight | Acute mammalian toxicity (proxy for people). |
| WHO class | Ia, Ib, II, III, U | WHO Recommended Classification of Pesticides by Hazard. |
| Fish LC50, Daphnia EC50, algae/Lemna EC50 | mg/L | Acute aquatic toxicity, 96 h, 48 h and 72 h–7 d respectively. |
| Honeybee LD50 | µg per bee | Acute contact toxicity. |
| Bird LD50 | mg/kg body weight | Acute oral toxicity, most sensitive species reported. |
| Earthworm LC50 | mg/kg soil | 14 d acute toxicity to Eisenia. |
| Typical rate | g active ingredient/ha | A representative single application rate, editable in the plan. |
| Withholding periods | days | Indicative grazing and harvest WHPs, with a text note. |
| NZ status | — | Approved, restricted, under reassessment, being phased out, or revoked. |
A value written >5000 is a limit test: the true figure is higher. Missing values are shown as a dash and handled as data gaps (§10). Biological products (Bt, granulovirus, Serratia, Trichoderma and others) have no meaningful Koc or DT50 and are scored as negligible for the fate-based components. Records marked ≈ carry particularly uncertain values.
3The situation
The Situation tab collects the land use and target (which select the candidate actives), the delivery method, the soil (organic carbon, bulk density ρb, water content at field capacity θ, macropore bypass fraction fp), the mean soil temperature, depth to groundwater and annual drainage, the distance and depth of the nearest water body, slope, wind and drainage, the sensitive receptors, and the user's constraints and weights. Soil defaults (in D.soils) are indicative NZ topsoil values and should be replaced with measured ones.
4Hazard classes
Each property is put into one of five classes, from 1 (very low hazard) to 5 (very high). The colour tints in the app always appear with the number or a word, never alone.
| Property | 1 very low | 2 low | 3 moderate | 4 high | 5 very high | Basis |
|---|---|---|---|---|---|---|
| Koc (mL/g) | > 5000 | 2000–5000 | 500–2000 | 150–500 | < 150 | McCall et al. (1981) mobility |
| Site Kd (L/kg) | > 100 | 20–100 | 5–20 | 1–5 | < 1 | KiwiPest |
| Soil DT50 (d) | ≤ 10 | 10–30 | 30–100 | 100–365 | > 365 | PPDB persistence |
| GUS | ≤ 1.0 | 1.0–1.8 | 1.8–2.8 | 2.8–4.0 | > 4.0 | Gustafson (1989) |
| Rat LD50 (mg/kg) | > 5000 | 2000–5000 | 300–2000 | 50–300 | ≤ 50 | GHS acute oral categories |
| Aquatic EC/LC50 (mg/L) | > 100 | 10–100 | 1–10 | 0.1–1 | ≤ 0.1 | GHS aquatic acute |
| Bee LD50 (µg/bee) | > 100 | 11–100 | 2–11 | 0.1–2 | ≤ 0.1 | US EPA bee categories |
| Bird LD50 (mg/kg) | > 2000 | 500–2000 | 50–500 | 10–50 | ≤ 10 | US EPA avian categories |
| Earthworm LC50 (mg/kg) | > 1000 | 500–1000 | 100–500 | 10–100 | ≤ 10 | PPDB |
| Solubility (mg/L) | ≤ 1 | 1–50 | 50–500 | 500–10 000 | > 10 000 | PPDB |
| WHO class | U | — | III | II | Ia, Ib | WHO (2019) |
5Sorption and Kd
Most organic pesticides sorb mainly to soil organic matter, so their distribution coefficient between soil and water scales with the organic carbon content:
Kd = Koc × foc, where foc = OC% / 100.
Kd in L/kg (mL/g). A silt loam with 3.5 % OC gives foc = 0.035, so picloram (Koc 35) has Kd = 1.2 L/kg and glyphosate (Koc 1424) has Kd = 50 L/kg.
The relation holds poorly for actives that sorb by other mechanisms: ionic species such as paraquat, diquat and glyphosate bind to clays and oxides, and acidic herbicides sorb less at high pH. NZ Allophanic soils sorb phosphonates and some anions strongly through their allophane and ferrihydrite. For copper, a tabulated Kd (1000 L/kg; the literature range is 50–30 000) is used directly.
6Degradation and temperature
Tabulated half-lives apply at about 20 °C. Most NZ soils are cooler, which slows microbial degradation. KiwiPest uses the Arrhenius-equivalent Q10 correction adopted by EFSA (2007):
DT50(T) = DT50(20 °C) × Q10(20 − T)/10, with Q10 = 2.58.
At a mean soil temperature of 12 °C the factor is 2.580.8 = 2.13, so a 30-day half-life becomes 64 days.
Soil moisture also affects degradation; KiwiPest does not correct for it.
7Leaching to groundwater
Groundwater ubiquity score
Gustafson's (1989) empirical index separates actives found in groundwater from those that are not:
GUS = log10(DT50) × (4 − log10 Koc)
GUS > 2.8: probable leacher. 1.8–2.8: transitional. < 1.8: improbable leacher.
KiwiPest reports GUS both at 20 °C (as in the literature) and with the temperature-corrected DT50. The Compare tab plots every active on these axes.
Attenuation factor
GUS ignores the site. For that, KiwiPest uses the attenuation factor of Rao, Hornsby and Jessup (1985) — the fraction of the applied active that survives the journey to the water table, assuming first-order degradation during advective transport:
R = 1 + ρb Kd / θ (retardation factor)
t = L R θ / q (travel time, days), with L the depth to groundwater (m) and q the drainage flux (m/day = mm/yr ÷ 1000 ÷ 365)
AF = exp(−0.693 t / DT50(T))
Applied to the whole soil water, this "piston flow" form predicts that almost nothing leaches more than a metre or two, which is contrary to the many detections of atrazine, terbuthylazine, hexazinone, picloram and bentazone in NZ groundwater. Field leaching is dominated by preferential flow through cracks, root channels and worm burrows. KiwiPest therefore routes a fraction fp of the drainage through a macropore pathway with a small mobile water content (θm = 0.05) that meets only a tenth of the sorption sites:
Rbp = 1 + 0.1 ρb Kd / θ tbp = L Rbp θm / q
AF = (1 − fp) exp(−0.693 t / DT50) + fp exp(−0.693 tbp / DT50)
Default fp runs from 0.02 in sands to 0.15 in structured clays (D.soils). For copper, which does not degrade, the groundwater score falls from 1 at a travel time of a year to 0 at a century.
The groundwater score is (log10 AF + 6) / 5, clipped to 0–1: AF = 0.1 scores 1, AF = 10−3 scores 0.6, and AF ≤ 10−6 scores 0. It is scaled by the share of the rate reaching the soil for the method (granules 1, boom 0.8, airblast 0.5, baits 0.1).
8Surface water
Spray drift
Drift deposited on water at distance d metres from the edge of the treated area, as a percentage of the rate, follows a power law fitted to the 90th-percentile drift tables used in European exposure assessment (Rautmann et al. 2001; FOCUS 2001):
drift% = A × d−b
Ground boom A = 2.77, b = 0.98 (0.57 % at 5 m, 0.15 % at 20 m) · orchard airblast A = 73.3, b = 1.40 · vineyard/kiwifruit airblast A = 20, b = 1.25 · aerial A = 200, b = 1.0 (indicative) · knapsack A = 0.5, b = 1.0. Baits, granules, seed treatments, targeted stem and stump treatments and fumigants have no spray drift.
Drift is scaled by wind speed ÷ 10 km/h (bounded 0.3–3). Where the land use is waterways and the active is used on aquatic weeds, the whole rate is taken as applied to the water.
Runoff
The runoff loss reaching the water body is an index rather than a hydrological prediction:
fro = rslope × krain × kdrain × (1/R + 0.02) × exp(−0.693 × 3 / DT50) × exp(−d / 50) × 10 × fsoil
rslope = 0.002 (flat), 0.01 (rolling), 0.03 (steep); krain = 3 if heavy rain is forecast; kdrain = 2 for mole or tile drains; 1/R is the dissolved share and 0.02 a sorbed share carried on sediment; the first rain is assumed three days after application; runoff falls off with buffer distance d; and ten units of treated area drain to each unit of water.
Concentration and risk
PEC (mg/L) = rate (kg/ha) × (drift fraction + fro) / (10 × water depth, m)
1 kg/ha deposited on 30 cm of water gives 0.33 mg/L.
ratio = max over fish, Daphnia and algae of PEC × trigger / EC50, with the acute triggers of the EU uniform principles: 100 for fish and invertebrates, 10 for algae and aquatic plants. A ratio ≥ 1 fails.
surface-water score = (log10 ratio + 2) / 3, clipped to 0–1 (a ratio of 0.01 scores 0; 1 scores 0.67; 10 scores 1).
For baits and other actives without a rate, the score is the aquatic hazard class scaled to 30 %.
9Bees, soil life, wildlife and people
Pollinators
HQ = rate (g ai/ha) / contact LD50 (µg/bee) (EPPO 2010; HQ > 50 triggers further assessment)
bee score = (log10(HQ × exposure) − 1) / 2.5, clipped to 0–1, where exposure is the method factor (spray 1, knapsack 0.5, seed treatment 0.3, baits, granules and targeted treatments 0.05) × 0.4 unless bees are foraging. Systemic actives add 0.15 when bees are foraging.
Soil life and persistence
PECsoil (mg/kg) = rate (kg/ha) × fsoil × 2 / ρb (mixed into the top 5 cm)
Toxicity score = (log10(10 × PECsoil / LC50worm) + 2) / 3; persistence score rises from 0 at DT50(T) = 30 d to 1 at 1000 d on a log scale. The soil score is the larger of the two. Copper, which never degrades, scores 1 for persistence.
Birds and wildlife
The bird hazard class (§4), as (class − 1)/4, is scaled by the dietary exposure of the method (bait, granule and seed treatment 1; spray 0.6; knapsack 0.3; targeted 0.05). Vertebrate toxic agents add 0.3 for secondary poisoning, 0.45 for second-generation anticoagulants; other actives with log Kow above 5 add 0.1. Where a vertebrate toxic agent has no bird LD50, its mammalian class (× 0.8) stands in for non-target mammals.
People
The larger of the rat LD50 class and the WHO class, as (class − 1)/4. Fumigants add 0.2 for inhalation exposure. This is a hazard proxy; operator exposure depends on formulation, PPE and method, which KiwiPest does not model.
10The risk index
Index = 100 × Σ wi mi si / Σ wi mi
si is the score (0–1) of each of the six components, wi the user's weight (0–3, default 1) and mi a site multiplier: groundwater × 1.5 with a nearby bore or spring; surface water × 1.5 for a sensitive waterway; pollinators × 2 when bees forage; wildlife × 1.5 beside native habitat; people × 1.5 with public access.
A component with no data is left out and the remaining weights are renormalised; the detail view lists the gaps. The index ranks options against each other for one situation. It is not an absolute measure of acceptability, and two options a few points apart are not meaningfully different.
11Flags and set-asides
An option is set aside (listed below the ranking, if shown at all) when its approval has been revoked, when the block is certified organic and the active is not permitted, or when its indicative WHP is longer than the days available before harvest or grazing. Other conditions raise flags without excluding the option:
- being phased out, under reassessment or restricted;
- not approved in the EU (for export produce);
- a mode-of-action group already used this season;
- bee toxicity, and systemic movement into pollen and nectar;
- an attenuation factor above 10−5 (caution) or 10−3 (limit), or a GUS above 2.8 near a water supply;
- an aquatic ratio ≥ 1, or very high aquatic toxicity beside a sensitive waterway;
- rain forecast for a mobile active; wind above 15 km/h or calm inversion conditions;
- hormone (HRAC 4) herbicide drift onto susceptible crops;
- broad-spectrum insecticides (IRAC 1A, 1B, 3A) disrupting biological control;
- persistence and copper accumulation; bioaccumulation potential;
- dogs, livestock and caution periods for vertebrate toxic agents;
- very high acute human toxicity, and public access.
12Worked example
Picloram for gorse on a rolling silt loam pasture: OC 3.5 %, ρb 1.2 g/cm³, θ 0.32, fp 0.06, soil 12 °C, groundwater at 5 m, drainage 300 mm/yr, stream 20 m away and 0.3 m deep, ground boom at 300 g ai/ha, wind 10 km/h.
| Step | Calculation | Result |
|---|---|---|
| Kd | 35 × 0.035 | 1.23 L/kg |
| DT50 at 12 °C | 82 × 2.580.8 = 82 × 2.13 | 175 d |
| GUS at 20 °C / at site | log 82 × (4 − log 35) / log 175 × (4 − log 35) | 4.70 / 5.51 |
| R, Rbp | 1 + 1.2 × 1.23 / 0.32; 1 + 0.1 × 4.59 | 5.59; 1.46 |
| q | 0.3 m / 365 d | 8.2 × 10−4 m/d |
| Travel time, matrix / macropore | 5 × 5.59 × 0.32 / q; 5 × 1.46 × 0.05 / q | 29.8 yr / 444 d |
| AF | 0.94 × e−43 + 0.06 × e−1.76 | 1.0 × 10−2 — limit flag |
| Groundwater score | (log 0.0103 + 6) / 5 | 0.80 |
| Drift at 20 m | 2.77 × 20−0.98 | 0.147 % |
| Runoff fraction | 0.01 × (1/5.59 + 0.02) × e−0.012 × e−0.4 × 10 × 0.8 | 0.0105 |
| PEC | 0.3 × (0.00147 + 0.0105) / 3 | 1.2 µg/L |
| Aquatic ratio (fish) | 0.0012 × 100 / 19.3 | 0.0062 → score 0 |
| Soil persistence | (log 175 − log 30) / (log 1000 − log 30) | 0.50 |
| People | LD50 4012 → class 2 | 0.25 |
| Index | 100 × (0.80 + 0 + 0 + 0.50 + 0 + 0.25) / 6 | 26 |
Triclopyr at 1200 g/ha in the same paddock indexes 23: it is about 35 times less likely to reach the water table (AF 2.8 × 10−4) and less persistent, but more toxic to people and applied at a higher rate. The difference is mainly groundwater — which is the point the comparison should make to the person choosing.
13Limitations
- Data. Single typical values hide wide variation between soils, studies and formulations. Some records (marked ≈) are approximate. Formulation co-formulants, surfactants and esters can be far more toxic than the active ingredient (glyphosate with tallow amine; triclopyr and 2,4-D esters).
- Metabolites are not modelled, although several are the real concern (AMPA, ETU from mancozeb, BAM from dichlobenil, desthio-prothioconazole, TFA from flufenacet, carbendazim from thiophanate-methyl).
- Chronic and sub-lethal effects — endocrine disruption, carcinogenicity, developmental neurotoxicity, chronic toxicity to aquatic invertebrates and bee colonies — are not scored, only mentioned in notes.
- Fate uses first-order kinetics, a single soil layer and a steady drainage flux. Subsoil degradation is usually slower than topsoil, which the model does not represent.
- Runoff and drift are indices, not predictions of concentration.
- Status and WHPs change. Check the EPA's register and the ACVM register before relying on any status shown.
- Efficacy and resistance are not scored beyond the mode-of-action check.
14Maintaining the data
All data are plain JavaScript files, edited by hand:
data/actives.js— one record per active; the key list is at the top of the file. Add a record, give it a newid(never reuse or change an existing one), and list its sectors and targets.data/reference.js— sectors, targets and their IPM notes, soils, application methods and drift curves, hazard class edges, status codes.js/model.js— the equations; constants are inKP.CONSTANTS. A change here needs the matching change on this page.
After editing, open the tool and run KP.selfTest() in the browser console: it checks every reference between the files and that every index falls within 0–100.
15References
- EFSA (2007). Opinion on a request from EFSA related to the default Q10 value used to describe the temperature effect on transformation rates of pesticides in soil. EFSA Journal 622: 1–32.
- EPPO (2010). Environmental risk assessment scheme for plant protection products, Chapter 10: honeybees. EPPO Bulletin 40: 323–331.
- FOCUS (2001). FOCUS Surface Water Scenarios in the EU Evaluation Process under 91/414/EEC. SANCO/4802/2001-rev.2.
- Gustafson, D.I. (1989). Groundwater ubiquity score: a simple method for assessing pesticide leachability. Environmental Toxicology and Chemistry 8: 339–357.
- Lewis, K.A., Tzilivakis, J., Warner, D. & Green, A. (2016). An international database for pesticide risk assessments and management. Human and Ecological Risk Assessment 22: 1050–1064. (PPDB)
- McCall, P.J., Laskowski, D.A., Swann, R.L. & Dishburger, H.J. (1981). Measurement of sorption coefficients of organic chemicals and their use in environmental fate analysis. In: Test Protocols for Environmental Fate and Movement of Toxicants. AOAC, Arlington, 89–109.
- Rao, P.S.C., Hornsby, A.G. & Jessup, R.E. (1985). Indices for ranking the potential for pesticide contamination of groundwater. Soil and Crop Science Society of Florida Proceedings 44: 1–8.
- Rautmann, D., Streloke, M. & Winkler, R. (2001). New basic drift values in the authorization procedure for plant protection products. Mitteilungen aus der Biologischen Bundesanstalt 383: 133–141.
- Standards New Zealand (2021). NZS 8409:2021 Management of agrichemicals.
- WHO (2020). The WHO Recommended Classification of Pesticides by Hazard and Guidelines to Classification 2019. World Health Organization, Geneva.
To use the tool rather than read about it, see the User guide. Where each number comes from is set out in Data sources & terms.