Overview of Project
Environmental Factors Contributing to Cryptosporidium Incidence in Massachusetts
This project proposes to analyze the spatial relationship between cryptosporidiosis incidence and dairy farm locations in Massachusetts. Cryptosporidiosis is a disease transmitted by the parasite, Cryptosporidium, which is commonly found in the manure of livestock and wild animals. There are two species that are transmittable to humans, C. parvum and C. hominis. C. parvum is largely found on dairy farms, where young calves are commonly infected. C. hominis is found among human populations and is transmitted between people. Prior outbreaks of cryptosporidiosis have been traced to dairy farms, due to an abundance of manure that can easily run-off into surface waters. Although Massachusetts has not had a serious outbreak, a spike in incidence did occur in Worcester a few years ago. The Massachusetts department of Public Health gathers passive surveillance data on cryptosporidiosis incidence from health practitioners who report the disease to the public health department. Incidence data has been made available to me by zipcode.
In addition to locations of dairy farms, I will also attempt to map dairy farm size and to relate this to incidence of disease. I have a map of dairy farm sizes and locations in Massachusetts that needs to be geo-referenced using the Massachusetts state plane data. Other elements in the analyses include precipitation data, hydrological features, slope and soil type, proximity to residential areas, proximity to public water supplies (this may not be available), proximity to surface water and land use characteristics. I plan to use MassGIS data layers corresponding to these elements.
Using overlay and zonal analysis features I will describe the pattern between cryptosporidiosis incidence and a multitude of environmental factors. I may experiment with including beef operations in addition to dairy operations. I may also attempt to use precipitation, slope, soil type, and proximity to a watercourse to calculate a run-off potential index like the one described in the study below. Then I could join the ranked RPI areas to diary farm areas. The dairy farm areas could also be ranked depending on their size and proximity to surface waters. I could also assign these areas an average cryptosporidium oocyst production value (see study below). Perhaps I could compare the run-off potential with the actual incidence data. Finally, the cryptosporidiosis incidence data will be joined to the zipcode locations of dairy farms. The points in this dataset might be related to the town centroid within a zipcode because this data has been used with GIS software for prior analyses.
Data Layer Source Accuracy
Precipitation data USGS Not sure
Soil type MassGIS 30 ft
Hydrology MassGIS 30 ft
Dairy Farm Locations Map (in my possession) 30 ft
Cryptosporidiosis Incidence MDPH Zipcode
Dairy Farm Size (# of head) Map (in my possession) 30 ft
Land Use MassGIS 30 ft
Public Water Supply intakes MassGIS (possibly, may not be available) 30 ft
Locations of private wells May not be available 30 ft
# of head of Cattle by zipcode U.S. Census Zipcode
Several studies using GIS in risk-analysis for cryptosporidiosis infection have helped me to formulate this idea:
Foster, J.A., McDonald, A.T. 2000. Assessing pollution risks to water supply intakes using geographical information systems (GIS). Environmental Modeling and Software. Vol 15, Issue 3. P 225-234.
This article described the use of GIS for risk assessment or cryptosporidium run-off overland in a catchment area in upland England. Researchers created a Runoff Potential Index (RPI) by combining the annual effective rainfall, land slope, soil hydrology and proximity to a watercourse. The ordinal ranked data were combined using geographic overlay, with calculations based on the smallest spatial unit (i.e. the slope - created from a triangular irregular network (TIN) in ArcInfo).
The RPI was used to identify areas at risk of contamination by cryptosporidium. To identify sources of contamination, animal stock numbers from agricultural census data were multiplied by average animal infection and faecal production rates to determine the predicted cryptosporidium loading for each grid cell in the area. To reduce spatial approximations made from census data, the Institute of Terrestrial Ecology (ITE) Landcover data map was introduced to identify where stock were located. The 25 m Landcover cells where livestock was found were assigned an average oocyst production value for cryptosporidium from the 2 km data. This means that only those areas where the source could be found (i.e. where animals were present) were displayed on the resultant hazard map.
The previously calculated RPI cells were combined with the cryptosporidium oocyst loaded cells by joining tables and selecting all RPI values contained within the load ranked cells. The resultant table contained attributes describing the runoff potential and cryptosporidium load, expressed as a rank. Cells were then displayed on a thematic map to illustrate areas of greatest risk to water quality from cryptosporidium.
2. Graczyk, Thaddeus K., Evans, B., Shiff, C., Karreman, H., Patz, J. 2000. Environmental and Geographical Factors Contributing to Watershed Contamination with Cryptosporidium parvum Oocysts. Environmental Research Section A 82, 263-271.
GIS software used: ARC/INFO, Arc/View
GIS was used to provide spatial analyses for the study, characterizing environmental and geographical factors contributing to watershed contamination with Cryptosporidium parvum oocysts. GIS data sets included a map layer depicting 100-year floodplain boundaries in the Pequea Creek watershed. A second layer depicted detailed land use and cover (including the location of livestock operations, cultivated lands, and grazing areas). The third, a "livestock population" layer, provided information on the type and number of domestic animals (e.g., dairy and beef cattle, hogs, sheep, horses, and chickens) by U.S. postal zip code boundary. The floodplain layer was created by digitizing the 100-year floodplain boundaries as represented on flood insurance rate maps for the three townships in the study area. Land cover/land use information was developed by interpretation of aerial photographs over the study area in 1995. This information was digitized using ARC/INFO software. The livestock population layer used in the study was developed by the U.S. Bureau of the Census as part of its regular national agricultural census activities.
3. Mc Donald, S., Murphy, T., Holden, N. 2007. Spatial and Temporal Issues in the Development of a Microbial Risk Assessment Model for Cryptosporidium. Making Science Work on the Farm: A Workshop on Decision Support Systems for Irish Agriculture.
In Ireland, GIS is also being used to develop a microbial risk assessment model to identify when potable water is at high risk of being contaminated with zoonotic enteric pathogens via surface pathways. GIS spatial data layers include: soils; geology; stock density; animal husbandry; farm storage of animal wastes; animal waste disposal practices; data from biological surveillance of the environment; and meteorological data. The overview of the ongoing study was not specific as to how these variables would be integrated. The study mentioned that ongoing sampling took place at 50% of the farms in the catchment area. Soil samples were taken from lands with different agricultural uses such as grazing, slurry spreading, silage, forestry, and calf paddocks. It was stressed that the GIS model being developed will have the capacity to factor in risk over time, and that this is more meaningful than current static models because the farming environment is constantly changing.
4. Lake, Iain, R. et al. 2007. Case-control study of environmental and social factors influencing cryptosporidiosis. European Journal of Epidemiology. 22:805-811.
This study investigated the role of wider environmental and socioeconomic factors upon human cryptosporidiosis. Using GIS, the detailed locations of 3368 laboratory-confirmed cases were compared to the locations of an equal number of controls. The cases were genotyped according to one of two types of species (C. parvum or C. hominis) to be examined separately. When all cryptosporidiosis cases were analyzed several location variables were strongly associated with illness: areas with higher socioeconomic status individuals, many individuals aged less than 4 years, areas with a high estimate of Cryptosporidium applied to land from manure, and areas with poorer water treatment.
Based on a 1 Km2 map of manure applications developed by the Agricultural Land Advisory Service, GIS was used to extract estimates of the total amount of Cryptosporidium applied to land through animal manures. Information from the agricultural census was combined with information from animal excreta, manure management surveys, and estimates of oocyst concentrations in manure.
Socioeconomic variables were obtained by identifying the 2001 Output area within which each postcode was located. Output areas are the smallest unit of the Census and contain approx. 125 people. Within an Output area, the percentage of people in eight socio-economic status bands was identified.
Information on the proportion of water supplied from different sources (surface vs. groundwater) and subject to different treatments (e.g. membrane filtration, simple disinfection) was obtained from the England and Wales Drinking Water Inspectorate. This information is only applicable to public water supplies. For surface or groundwater abstraction, catchments were calculated using GIS and the density of Cryptosporidium applications to land, sewage discharges and sewage overflows in each catchment were calculated.
Over 50 explanatory variables were produced. Multivariate models were created using logistic regression. Significant positive risk factors for Cryptosporidiosis included: living in an area with higher amounts of Cryptosporidium applied to land in a 2.5 km buffer around each postcode (OR 1.084 P =0.02), larger proportions of individuals in the 0-4 years age group (OR 1.145 P<0.001) and more individuals in the highest socioeconomic status groups (OR 1.203 P<0.001). Factors negatively affecting risk were: Drinking water subject to superior treatment (OR 0.770 P<0.001) and groundwater sourced drinking water (OR 0.821 P=0.001).