Summary and conclusions, Miljøstyrelsen

Risikovurdering ved anvendelse af vandingskanoner til udspredning af gylle fortyndet med vand

Summary and conclusions

The present report summarise a risk assessment of applying big volume irrigation guns to spread manure/slurry diluted with water. The report describes human health risks directly associated with exposure to aerosols originating from the irrigation.

Authorities has been precarious about the risk of airborne transmitted pathogenic bacteria, virus and protozoan (agens), that might arise from applying irrigation guns to spread manure/slurry diluted with water, and implemented the present risk assessment of the probability of human exposure to aerosols. Aerosols from the irrigation may contain pathogenic agens and will, depending on the size of the aerosols and the climatic conditions, drift over greater or smaller distances, thereby posing a risk of causing airborne infections of humans and animals. A number of factors are important for the drift of aerosols. Most important is wind-movement; furthermore temperature, relative humidity and atmospherically stability are important factors for the possible drift of generated aerosols.

After a brief introduction in Chapter 1, the most relevant zoonotic microorganisms present in manure/slurry from Danish livestock are described in Chapter 2 (zoonotic microorganisms are diseases and infections transmitted from animals to humans. The bacteria Campylobacter, Salmonella and Yersinia cause the majority of zoonotic infections in Denmark). The description includes source of infection, dissemination, dose of infection, course and symptoms of the disease, occurrence in Danish livestock, occurrence in manure/slurry and decimation time in manure/slurry. The literature is sparse on quantitative description of the microbial composition in a representative section of Danish livestock.

Salmonella and Cryptosporidia are chosen for the quantitative risk assessment Based on two criteria: 1) Data is available on occurrence and survival in manure/slurry (Salmonella up to 2,8*10⁴ and Cryptosporidium up to 3*10² per ml. in undiluted manure/slurry), and 2) both are relevant under the existing conditions in Denmark, i.e. importance and frequency (pose a risk) in Danish livestock and for humans.

Chapter 3 concise describe the techniques applied to spread manure/slurry. Today about 75 % is spread using "drag tubes"; about 15 % is spread using a traditional spreading machine, and about 10 % by burying manure. Especially spread of manure/slurry with irrigation guns is outlined since the report primarily deals with and assesses this method.

Chapter 4 gives a brief introduction to aerosols and their drift potential (primarily drop size). It is only relevant to look at drops with a diameter smaller than 1 mm. since larger drops, under normal wind conditions, fall quickly to the ground and therefore the drift is minimal. The dropvolumedistribution is described based on dimensions of an irrigation system for spread of manure diluted with water. The distribution describes the proportion of the total volume found in drops with a diameter smaller than a given size. 20 % of the total volume is found in drops with a diameter less than 1 mm.

Chapter 5 describe the movement and drift of aerosols in open air. Modelling of aerosol drift is done using a special developed version of the model RIMPUFF (RIsø Mesoscale PUFFmodel). RIMPUFF is suitable for predicting the spread of material (aerosols, particles etc.) in the atmosphere, where the meteorological parameters vary in time and space. The result of the drift modelling is described on a 5x5 km. grid with a grid size of 20x20m and includes the amount of diluted manure/slurry per m³ one meter above the ground and the amount of diluted manure/slurry deposited per m² on the ground. The uncertainty of the model is evaluated several times and it predicts within a factor of 2 to 3.

Chapter 6 consist of a qualitative assessment of the human and animal health risks caused by staying in an area affected by the aerosol cloud Focus is on aerogene (airborne contact) and peroral (through the mouth, direct contact) human infections from the aerosol cloud. Very few data on human infective dose for aerogene infections exists. However, other infections than the traditional airborne can spread in this manner and the infective dose is supposed to be lower compared to peroral infections. People staying in an area affected by the aerosol cloud might be exposed to agens, resulting in aerogene or peroral infections from the aerosol cloud.

In case people stay in the aerosol affected area contamination of foodstuff, clothing, skin or brought objects is possible, thereby posing a risk for subsequent peroral infection (through the mouth, indirect contact).

Animals staying in the aerosol-affected area may be exposed to agens resulting in aerogene (airborne contact) or peroral (through the mouth, direct contact) infections from the aerosol cloud. Furthermore, animals grazing on areas or animals that later in time are fed with or in contact with hay/straw exposed to the aerosol cloud, might risk infections from deposited agens (indirect contact).

Furthermore, the risk of spreading antimicrobial-resistant organisms to the environment and the possible extra consequences of having an infection with antimicrobial-resistant bacteria should be considered.

Finally, chemical substances such as, remnants of medicine, hormones, mercaptanes, carbondioxid, ammonia, methane, disinfectants, hydrogen sulphide etc. might be spread along with the manure. The hazard of these substances are not assessed in this report.

The following risk reducing actions will decrease the probability of human infections with agens from the aerosol cloud.

Storage of the manure or other treatment that reduces the infectious matter.
Only spread in calm wind.
Apply techniques that minimize generation of aerosols.

Safety precautions concerning spread of manure/slurry must be accommodated to the relevant risks. More stringent measures must be taken at livestock that pose a high risk of spreading agens, i.e. it is prohibited to spread manure/slurry from livestock with clinical salmonellosis or livestock under public supervision. This is relevant concerning transfer of agens to both humans and animals.

Chapter 7 comprise a quantitative risk assessment, including Salmonella and Cryptosporidia. The assessment results in two measures for the spread of zoonotic agens from the irrigation system.

Number of agens inhaled per hours at a given location in the area affected by the aerosol cloud (outside the sprayfield).
Number of agens deposited per m² during the spray period at a given location in the area affected by the aerosol cloud (outside the sprayfield).

The results are described by means of different scenarios and Monte Carlo simulations. In Monte Carlo simulations the probability distributions of the uncertain elements are used as a basis for sampling to simulate a distribution of the outcome (here number of agens inhaled per hours or the number of agens deposited per m² during the spray period).

Salmonella, worst-case scenario

The number of bacteria in the undiluted manure/slurry is assumed to be 2.8´ 10⁴(found in the Danish investigation). Furthermore, it is assumed that the bacteria is fully mixed in the manure and the proportion attached to particles can be neglected. All bacteria are assumed to survive in the aerosol cloud. The proportion of manure is 25%. The table below show worst-case estimates, the upper row is the distance from the property-line of the sprayfield (in meters) and the lower row is the number of Salmonella bacteria inhaled per hour having a strong respiration (30 m³/day ť 1250 litre/hour).

Distance (m)	0	50	300	500	800	2000	5000
Number/hour	~10000	~4000	~900	~500	~300	~100	~30

The number of Salmonella bacteria inhaled decreases with increasing distance from the property-line of the sprayfield. At the property-line of the sprayfield approximately 10⁴ Salmonella bacteria is inhaled per hour and bacteria is presumably inhaled at a distance of 5 km. from the sprayfield (in the wind direction).

The table below show worst-case estimates, the upper row is the distance from the property-line of the sprayfield (in meters) and the lower row is the number of Salmonella bacteria deposited per m² during the entire spray period (8 hours).

Distance (m)	0	50	300	500	800	2000	5000
Number/m²	1*10⁷	~200000	~8000	~700	~150	~65	~13

The number of Salmonella bacteria deposited decreases with increasing distance from the property-line of the sprayfield. At the property-line of the sprayfield approximately 10⁷ Salmonella bacteria is deposited per m² during the entire spray period (8 hours) and bacteria is presumably deposited at a distance of 5 km. from the sprayfield (in the wind direction).

Salmonella, Monte Carlo simulations

The number of bacteria in the undiluted manure/slurry is described by the distribution shown in Figure 7. It is assumed that the proportion of bacteria attached to particles varies uniformly between 25 % and 75 %. Furthermore the parameters in the drop volume-distribution are varied with a gaussian distribution with mean at the experimentally found values and a standard deviation of 10% of the mean. Bacterial survival in the aerosol cloud varies uniformly between 50 % and 100%. The proportion of manure varies between 10 % and 25%. The respiratory volume is varied according to a gaussian distribution with mean of 25m³/day and a standard deviation of 3m³/day. The simulations used to evaluate the variation in the number of Salmonella bacteria inhaled per hour. The results from the simulations is shown in Figure 12, from which the quantiles/percentiles in the table below is extracted. The upper row is the distance from the property-line of the sprayfield (in meters) and the lower rows is the number of Salmonella bacteria inhaled per hour

Distance (m)	0	20	40	100	200	500	900
Median	1.3	0.8	0.3	0.11	0.07	0.02	0.015
5% lower	0.04	0.02	0.01	<0.01	<0.01	<0.01	<0.01
95% upper	36	22	11	3.5	2.2	0.8	0,5
Maximum	1970	1190	660	270	154	56	40

As expected the number of Salmonella bacteria inhaled per hour is less in the simulations compared to the worst-case scenario. To obtain the worst-case scenario in the Monte Carlo simulations the maximum value in all distributions must be drawn simultaneously.

Cryptosporidium, worst-case scenario

The number of Cryptosporidium in the undiluted manure/slurry is assumed to be 3´ 10². Furthermore, it is assumed that the Cryptosporidium is fully mixed in the manure and the proportion attached to particles can be neglected. All Cryptosporidium is assumed to survive in the aerosol cloud. The proportion of manure is 25%. The table below show worst-case estimates, the upper row is the distance from the property-line of the sprayfield (in meters) and the lower row is the number of Cryptosporidium inhaled per hour having a strong respiration (30 m³/day ť 1250 litre/hour)

Distance (m)	0	50	300	500	800	2000	5000
Number/hour	~139	~50	~9	~4	~2	~1	~0.4

The number of Cryptosporidium inhaled decreases with increasing distance from the property-line of the sprayfield. At the property-line of the sprayfield approximately 10² Cryptosporidium is inhaled per hour and Cryptosporidium is presumably inhaled at a distance of 2 km. from the sprayfield (in the wind direction).

The table below show worst-case estimates, the upper row is the distance from the property-line of the sprayfield (in meters) and the lower row is the number of Cryptosporidium deposited per m² during the entire spray period (8 hours)

Distance (m)	0	50	300	500	800	2000	5000
Number/m²	~50000	~7000	~100	~4	~1	~0.3	~0.3

The number of Cryptosporidium deposited decreases with increasing distance from the property-line of the sprayfield. At the property-line of the sprayfield approximately 10⁵ Cryptosporidium is deposited per m² during the entire spray period (8 hours) and Cryptosporidium is presumably deposited at a distance of 0.8 km. from the sprayfield (in the wind direction)

Cryptosporidium, Monte Carlo simulations

The number of Cryptosporidium in the undiluted manure/slurry is described by the distribution shown in Figure 13. It is assumed that the proportion of Cryptosporidium attached to particles varies uniformly between 25 % and 75 %. Furthermore the parameters in the drop volume-distribution are varied with a gaussian distribution with mean at the experimentally found values and a standard deviation of 10% of the mean. Cryptosporidium survival in the aerosol cloud varies uniformly between 50 % and 100%. The proportion of manure varies between 10 % and 25%. The respiratory volume is varied according to a gaussian distribution with mean of 25m³/day and a standard deviation of 3m³/day. The simulations used to evaluate the variation in the number of Cryptosporidium inhaled per hour the results from the simulations are shown in Figure 18, from which the quantiles/percentiles in the table below are extracted. The upper row is the distance from the property-line of the sprayfield (in meters) and the lower rows is the number of Cryptosporidium inhaled per hour

Distance (m)	0	20	40	100	200	500	900
Median	1.1	0.7	0.35	0.14	0.07	0.02	0.01
5% lower	0.28	0.17	0.07	0.03	0.02	<0.01	<0.01
95% upper	5	2	1,6	0.7	0.35	0.08	0.07
Maximum	20	15	8	4	2	0,5	0.4

As expected the number of Cryptosporidium inhaled per hour is less in the simulations compared to the worst-case scenario. To obtain the worst-case scenario in the Monte Carlo simulations the maximum value in all distributions must be drawn simultaneously.

Furthermore, a comparison of the risks associated with applying irrigation guns to spread manure/slurry with spread of manure/slurry using traditional techniques. It is concluded that the described techniques pose a substantial increase in human and animal risk compared to the traditional techniques.

Unsteady weather conditions will increase the spread of pathogens, since drops exposed to larger wind speed; high relative humidity, low temperatures and unstable meteorological conditions will drift longer.

Focus in the present risk assessment has been on zoonotic pathogens and human risk. There are though a number of aspects that should be mentioned when discussing the described technique. Manure/slurry may contain chemical substances such as, remnants of medicine, hormones, mercaptanes, carbondioxid, ammonia, methane, disinfectants, hydrogen sulphide etc. These substances all have a toxic potential and might be spread along with the manure. Based on the increased generation of aerosols it is assessed that the technique will increase the spread of these substances (possible chemical reactions in the air is not taken into account).

The veterinary aspects can be dealt with in the same manner as the human risk assessment; it only requires a scaling of the respiratory volume. Furthermore the number of agens deposited on the ground may be used to assess the risk of infecting farm animals grazing on (or animals eating hay from) pasture affected by the aerosol cloud. The number of agens deposited on the surrounding pastureland is increased applying the techniques described compared to the traditional techniques.