VectorNet: collaborative mapping of arthropod disease vectors in Europe and surrounding areas since 2010.

BackgroundArthropod vectors such as ticks, mosquitoes, sandflies and biting midges are of public and veterinary health significance because of the pathogens they can transmit. Understanding their distributions is a key means of assessing risk. VectorNet maps their distribution in the EU and surrounding areas.AimWe aim to describe the methodology underlying VectorNet maps, encourage standardisation and evaluate output.Methods: Vector distribution and surveillance activity data have been collected since 2010 from a combination of literature searches, field-survey data by entomologist volunteers via a network facilitated for each participating country and expert validation. Data were collated by VectorNet members and extensively validated during data entry and mapping processes.ResultsAs of 2021, the VectorNet archive consisted of ca 475,000 records relating to > 330 species. Maps for 42 species are routinely produced online at subnational administrative unit resolution. On VectorNet maps, there are relatively few areas where surveillance has been recorded but there are no distribution data. Comparison with other continental databases, namely the Global Biodiversity Information Facility and VectorBase show that VectorNet has 5-10 times as many records overall, although three species are better represented in the other databases. In addition, VectorNet maps show where species are absent. VectorNet's impact as assessed by citations (ca 60 per year) and web statistics (58,000 views) is substantial and its maps are widely used as reference material by professionals and the public.ConclusionVectorNet maps are the pre-eminent source of rigorously validated arthropod vector maps for Europe and its surrounding areas.

Seasonality and timing of peak abundance of Aedes albopictus in Europe: Implications to public and animal health.

Aedes albopictus is a known vector of dengue and chikungunya. Understanding the population dynamics characteristics of vector species is of pivotal importance to optimise surveillance and control activities, to estimate risk for pathogen-transmission, and thus to enhance support of public health decisions. In this paper we used a seasonal activity model to simulate the start (spring hatching) and end (autumn diapause) of the vector season. In parallel, the peak abundance of the species was assessed using both VectorNet field survey data complemented with field studies obtained from literature across the Mediterranean Basin. Our results suggest that spring hatching of eggs in the current distribution area can start at the beginning of March in southern Europe and in April in western Europe. In northern Europe, where the species is not (yet) present, spring hatching would occur from late April to late May. Aedes albopictus can remain active up to 41 weeks in southern Europe whilst the climatic conditions in northern Europe are limiting its potential activity to a maximum of 23 weeks. The peak of egg density is found during summer months from end of July until end of September. During these two months the climatic conditions for species development are optimal, which implies a higher risk for arbovirus transmission by Ae. albopictus and occurrence of epidemics.

Increased risk for autochthonous vector-borne infections transmitted by Aedes albopictus in continental Europe.

Autochthonous outbreaks of chikungunya and dengue during the past decade showed that continental Europe is vulnerable to Aedes albopictus-borne infections. Ae. albopictus has spread geographically, resulting in more people exposed to risk. Timely application of adequate mosquito suppression measures may delay, or even prevent, the vector population from crossing the potential epidemic abundance threshold should a pathogen be introduced. Health authorities should be on alert to detect early cases to prevent autochthonous outbreaks.

Effect of Climate and Land Use on the Spatio-Temporal Variability of Tick-Borne Bacteria in Europe.

The incidence of tick-borne diseases caused by Borrelia burgdorferi sensu lato, Anaplasma phagocytophilum and Rickettsia spp. has been rising in Europe in recent decades. Early pre-assessment of acarological hazard still represents a complex challenge. The aim of this study was to model Ixodes ricinus questing nymph density and its infection rate with B. burgdorferi s.l., A. phagocytophilum and Rickettsia spp. in five European countries (Italy, Germany, Czech Republic, Slovakia, Hungary) in various land cover types differing in use and anthropisation (agricultural, urban and natural) with climatic and environmental factors (Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Land Surface Temperature (LST) and precipitation). We show that the relative abundance of questing nymphs was significantly associated with climatic conditions, such as higher values of NDVI recorded in the sampling period, while no differences were observed among land use categories. However, the density of infected nymphs (DIN) also depended on the pathogen considered and land use. These results contribute to a better understanding of the variation in acarological hazard for Ixodes ricinus transmitted pathogens in Central Europe and provide the basis for more focused ecological studies aimed at assessing the effect of land use in different sites on tick-host pathogens interaction.

Forecasting the spatial and seasonal dynamic of Aedes albopictus oviposition activity in Albania and Balkan countries.

The increasing spread of the Asian tiger mosquito, Aedes albopictus, in Europe and US raises public health concern due to the species competence to transmit several exotic human arboviruses, among which dengue, chikungunya and Zika, and urges the development of suitable modeling approach to forecast the spatial and temporal distribution of the mosquito. Here we developed a dynamical species distribution modeling approach forecasting Ae. albopictus eggs abundance at high spatial (0.01 degree WGS84) and temporal (weekly) resolution over 10 Balkan countries, using temperature times series of Modis data products and altitude as input predictors. The model was satisfactorily calibrated and validated over Albania based observed eggs abundance data weekly monitored during three years. For a given week of the year, eggs abundance was mainly predicted by the number of eggs and the mean temperature recorded in the preceding weeks. That is, results are in agreement with the biological cycle of the mosquito, reflecting the effect temperature on eggs spawning, maturation and hatching. The model, seeded by initial egg values derived from a second model, was then used to forecast the spatial and temporal distribution of eggs abundance over the selected Balkan countries, weekly in 2011, 2012 and 2013. The present study is a baseline to develop an easy-handling forecasting model able to provide information useful for promoting active surveillance and possibly prevention of Ae. albopictus colonization in presently non-infested areas in the Balkans as well as in other temperate regions.

Current and future distribution of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) in WHO Eastern Mediterranean Region.

BACKGROUND: Aedes-borne diseases as dengue, zika, chikungunya and yellow fever are an emerging problem worldwide, being transmitted by Aedes aegypti and Aedes albopictus. Lack of up to date information about the distribution of Aedes species hampers surveillance and control. Global databases have been compiled but these did not capture data in the WHO Eastern Mediterranean Region (EMR), and any models built using these datasets fail to identify highly suitable areas where one or both species may occur. The first objective of this study was therefore to update the existing Ae. aegypti (Linnaeus, 1762) and Ae. albopictus (Skuse, 1895) compendia and the second objective was to generate species distribution models targeted to the EMR. A final objective was to engage the WHO points of contacts within the region to provide feedback and hence validate all model outputs. METHODS: The Ae. aegypti and Ae. albopictus compendia provided by Kraemer et al. (Sci Data 2:150035, 2015; Dryad Digit Repos, 2015) were used as starting points. These datasets were extended with more recent species and disease data. In the next step, these sets were filtered using the Köppen-Geiger classification and the Mahalanobis distance. The occurrence data were supplemented with pseudo-absence data as input to Random Forests. The resulting suitability and maximum risk of establishment maps were combined into hard-classified maps per country for expert validation. RESULTS: The EMR datasets consisted of 1995 presence locations for Ae. aegypti and 2868 presence locations for Ae. albopictus. The resulting suitability maps indicated that there exist areas with high suitability and/or maximum risk of establishment for these disease vectors in contrast with previous model output. Precipitation and host availability, expressed as population density and night-time lights, were the most important variables for Ae. aegypti. Host availability was the most important predictor in case of Ae. albopictus. Internal validation was assessed geographically. External validation showed high agreement between the predicted maps and the experts' extensive knowledge of the terrain. CONCLUSION: Maps of distribution and maximum risk of establishment were created for Ae. aegypti and Ae. albopictus for the WHO EMR. These region-specific maps highlighted data gaps and these gaps will be filled using targeted monitoring and surveillance. This will increase the awareness and preparedness of the different countries for Aedes borne diseases.

Climate-driven longitudinal trends in pasture-borne helminth infections of dairy cattle.

Helminth parasites of grazing ruminants are highly prevalent globally and impact negatively on animal productivity and food security. There is a growing concern that climate change increases helminth disease frequency and intensity. In Europe, these concerns stem from case reports and theoretical life cycle models assessing the effects of climate change scenarios on helminth epidemiology. We believe this study is the first to investigate climate-driven trends in helminth infections of cattle on a cohort of randomly selected farms. One thousand, six hundred and eighty dairy farms were monitored over an 8year period for the two major helminth infections in temperate climate regions and climate-driven trends were investigated by multivariable linear mixed models. The general levels of exposure to Fasciola hepatica decreased over the study period while those to Ostertagia ostertagi increased, and this could at least be partially explained by meteorological factors (i.e. the number of rainy (precipitation >1mm) and warm days (average daily temperature >10°C) in a year). The longitudinal trends varied according to the altitude and the agricultural region of the farm. This study shows that longitudinal epidemiological data from sentinel farms combined with meteorological datasets can significantly contribute to understanding the effects of climate on infectious disease dynamics. When local environmental conditions are taken into account, the effects of climate change on disease dynamics can also be understood at more local scales. We recommend setting up a longitudinal sampling strategy across Europe in order to monitor climate-driven changes in helminth disease risk to inform adaptation strategies to promote animal health and productivity.

The dog and cat population on Maio Island, Cape Verde: characterisation and prediction based on household survey and remotely sensed imagery.

The objective was to estimate and characterise the dog and cat population on Maio Island, Cape Verde. Remotely sensed imagery was used to document the number of houses across the island and a household survey was carried out in six administrative areas recording the location of each animal using a global positioning system instrument. Linear statistical models were applied to predict the dog and cat populations based on the number of houses found and according to various levels of data aggregation. In the surveyed localities, a total of 457 dogs and 306 cats were found. The majority of animals had owners and only a few had free access to outdoor activities. The estimated population size was 531 dogs [95% confidence interval (CI): 453-609] and 354 cats (95% CI: 275-431). Stray animals were not a concern on the island in contrast to the rest of the country.

Mapping and modelling helminth infections in ruminants in Europe: experience from GLOWORM.

Mapping and modelling helminth infections in cattle and sheep in Europe through advanced geospatial research was one of the main task of GLOWORM, a three year project (2012-2014) funded under the European Commission's (EC) seventh framework programme (FP7). Liver flukes as Fasciola hepatica and gastrointestinal nematodes, such as Haemonchus contortus were chosen for the project since these parasites constitute a major cause of lost productivity in small and large ruminants. The output of the GLOWORM project delivered guidelines for standardized and harmonized cross-sectional surveys of helminth parasites in ruminants allowing the development of updated prevalence maps and multi-scale, spatial models for the European area.

Modelling the spatial distribution of Fasciola hepatica in dairy cattle in Europe.

A harmonized sampling approach in combination with spatial modelling is required to update current knowledge of fasciolosis in dairy cattle in Europe. Within the scope of the EU project GLOWORM, samples from 3,359 randomly selected farms in 849 municipalities in Belgium, Germany, Ireland, Poland and Sweden were collected and their infection status assessed using an indirect bulk tank milk (BTM) enzyme-linked immunosorbent assay (ELISA). Dairy farms were considered exposed when the optical density ratio (ODR) exceeded the 0.3 cut-off. Two ensemble-modelling techniques, Random Forests (RF) and Boosted Regression Trees (BRT), were used to obtain the spatial distribution of the probability of exposure to Fasciola hepatica using remotely sensed environmental variables (1-km spatial resolution) and interpolated values from meteorological stations as predictors. The median ODRs amounted to 0.31, 0.12, 0.54, 0.25 and 0.44 for Belgium, Germany, Ireland, Poland and southern Sweden, respectively. Using the 0.3 threshold, 571 municipalities were categorized as positive and 429 as negative. RF was seen as capable of predicting the spatial distribution of exposure with an area under the receiver operation characteristic (ROC) curve (AUC) of 0.83 (0.96 for BRT). Both models identified rainfall and temperature as the most important factors for probability of exposure. Areas of high and low exposure were identified by both models, with BRT better at discriminating between low-probability and high-probability exposure; this model may therefore be more useful in practise. Given a harmonized sampling strategy, it should be possible to generate robust spatial models for fasciolosis in dairy cattle in Europe to be used as input for temporal models and for the detection of deviations in baseline probability. Further research is required for model output in areas outside the eco-climatic range investigated.

Spatial analysis and risk mapping of Fasciola hepatica infection in dairy herds in Ireland.

Fasciolosis is generally a subclinical infection of dairy cows and can cause marked economic losses. This study investigated the prevalence and spatial distribution of fasciolosis in dairy cow herds in Ireland using an in-house antibodydetection enzyme-linked immunosorbent assay applied to bulk tank milk (BTM) samples collected during the autumn of 2012. A total of 5,116 BTM samples were collected from 4,602 different herds, with 514 farmers submitting BTM samples in two consecutive months. Analysis of the BTM samples showed that 82% (n = 3,764) of the dairy herds had been exposed to Fasciola hepatica. A total of 108 variables, including averaged climatic data for the period 1981-2010 and contemporary meteorological data for the year 2012, such as soil, subsoil, land cover and habitat maps, were investigated for a possible role as predictor of fasciolosis. Using mainly climatic variables as the major predictors, a model of the predicted risk of fasciolosis was created by Random Forest modelling that had 95% sensitivity and 100% specificity. The most important predictors in descending order of importance were: average of annual total number of rain-days for the period 1981-2010, total rainfall during September, winter and autumn of 2012, average of annual total number of wet-days for the period 1981- 2010 and annual mean temperature of 2012. The findings of this study confirm the high prevalence of fasciolosis in Irish dairy herds and suggest that specific weather and environmental risk factors support a robust and precise distribution model.

Decision making on helminths in cattle: diagnostics, economics and human behaviour.

Helminth infections of cattle affect productivity in all classes of stock, and are amongst the most important production-limiting diseases of grazing ruminants. Over the last decade, there has been a shift in focus in the diagnosis of these infections from merely detecting presence/absence of infection towards detecting its impact on production. This has been facilitated by studies observing consistent negative correlations between helminth diagnostic test results and measures of productivity. Veterinarians are increasingly challenged to consider the economic aspects of their work, and the use of these tests should now be integrated in economic evaluation frameworks for improved decision making. In this paper, we review recent insights in the farm-specific economic impact of helminth infections on dairy cattle farms as well as in farmer attitudes and behaviour regarding helminth control. Combining better economic impact assessments of helminth infections together with a deeper understanding of the non-economic factors that drive a farmer's animal health decisions should result in more effective control strategies and increased farmer satisfaction.

Longitudinal study on the temporal and micro-spatial distribution of Galba truncatula in four farms in Belgium as a base for small-scale risk mapping of Fasciola hepatica.

BACKGROUND: The trematode parasite Fasciola hepatica causes important economic losses in ruminants worldwide. Current spatial distribution models do not provide sufficient detail to support farm-specific control strategies. A technology to reliably assess the spatial distribution of intermediate host snail habitats on farms would be a major step forward to this respect. The aim of this study was to conduct a longitudinal field survey in Flanders (Belgium) to (i) characterise suitable small water bodies (SWB) for Galba truncatula and (ii) describe the population dynamics of G. truncatula. METHODS: Four F. hepatica-infected farms from two distinct agricultural regions were examined for the abundance of G. truncatula from the beginning (April 2012) until the end (November 2012) of the grazing season. Per farm, 12 to 18 SWB were selected for monthly examination, using a 10 m transect analysis. Observations on G. truncatula abundance were coupled with meteorological and (micro-)environmental factors and the within-herd prevalence of F. hepatica using simple comparison or negative binomial regression models. RESULTS: A total of 54 examined SWB were classified as a pond, ditch, trench, furrow or moist area. G. truncatula abundance was significantly associated with SWB-type, region and total monthly precipitation, but not with monthly temperature. The clear differences in G. truncatula abundance between the 2 studied regions did not result in comparable differences in F. hepatica prevalence in the cattle. Exploration of the relationship of G. truncatula abundance with (micro)-environmental variables revealed a positive association with soil and water pH and the occurrence of Ranunculus sp. and a negative association with mowed pastures, water temperature and presence of reed-like plant species. CONCLUSIONS: Farm-level predictions of G. truncatula risk and subsequent risk for F. hepatica occurrence would require a rainfall, soil type (representing the agricultural region) and SWB layer in a geographic information system. While rainfall and soil type information is easily accessible, the recent advances in very high spatial resolution cameras carried on board of satellites, planes or drones should allow the delineation of SWBs in the future.

Optimizing denominator data estimation through a multimodel approach.

To assess the risk of (zoonotic) disease transmission in developing countries, decision makers generally rely on distribution estimates of animals from survey records or projections of historical enumeration results. Given the high cost of large-scale surveys, the sample size is often restricted and the accuracy of estimates is therefore low, especially when spatial high-resolution is applied. This study explores possibilities of improving the accuracy of livestock distribution maps without additional samples using spatial modelling based on regression tree forest models, developed using subsets of the Uganda 2008 Livestock Census data, and several covariates. The accuracy of these spatial models as well as the accuracy of an ensemble of a spatial model and direct estimate was compared to direct estimates and "true" livestock figures based on the entire dataset. The new approach is shown to effectively increase the livestock estimate accuracy (median relative error decrease of 0.166-0.037 for total sample sizes of 80-1,600 animals, respectively). This outcome suggests that the accuracy levels obtained with direct estimates can indeed be achieved with lower sample sizes and the multimodel approach presented here, indicating a more efficient use of financial resources.

A Bayesian geostatistical Moran Curve model for estimating net changes of tsetse populations in Zambia.

For the first time a Bayesian geostatistical version of the Moran Curve, a logarithmic form of the Ricker stock recruitment curve, is proposed that is able to give an estimate of net change in population demographic rates considering components such as fertility and density dependent and density independent mortalities. The method is applied to spatio-temporally referenced count data of tsetse flies obtained from fly-rounds. The model is a linear regression with three components: population rate of change estimated from the Moran curve, an explicit spatio-temporal covariance, and the observation error optimised within a Bayesian framework. The model was applied to the three main climate seasons of Zambia (rainy--January to April, cold-dry--May to August, and hot-dry--September to December) taking into account land surface temperature and (seasonally changing) cattle distribution. The model shows a maximum positive net change during the hot-dry season and a minimum between the rainy and cold-dry seasons. Density independent losses are correlated positively with day-time land surface temperature and negatively with night-time land surface temperature and cattle distribution. The inclusion of density dependent mortality increases considerably the goodness of fit of the model. Cross validation with an independent dataset taken from the same area resulted in a very accurate estimate of tsetse catches. In general, the overall framework provides an important tool for vector control and eradication by identifying vector population concentrations and local vector demographic rates. It can also be applied to the case of sustainable harvesting of natural populations.

Bluetongue, Schmallenberg - what is next? Culicoides-borne viral diseases in the 21st Century.

In the past decade, two pathogens transmitted by Culicoides biting midges (Diptera: Ceratopogonidae), bluetongue virus and Schmallenberg virus, have caused serious economic losses to the European livestock industry, most notably affecting sheep and cattle. These outbreaks of arboviral disease have highlighted large knowledge gaps on the biology and ecology of indigenous Culicoides species. With these research gaps in mind, and as a means of assessing what potential disease outbreaks to expect in the future, an international workshop was held in May 2013 at Wageningen University, The Netherlands. It brought together research groups from Belgium, France, Germany, Spain, Switzerland, United Kingdom and The Netherlands, with diverse backgrounds in vector ecology, epidemiology, entomology, virology, animal health, modelling, and genetics. Here, we report on the key findings of this workshop.

Fine-scale mapping of vector habitats using very high resolution satellite imagery: a liver fluke case-study.

The visualization of vector occurrence in space and time is an important aspect of studying vector-borne diseases. Detailed maps of possible vector habitats provide valuable information for the prediction of infection risk zones but are currently lacking for most parts of the world. Nonetheless, monitoring vector habitats from the finest scales up to farm level is of key importance to refine currently existing broad-scale infection risk models. Using Fasciola hepatica, a parasite liver fluke, as a case in point, this study illustrates the potential of very high resolution (VHR) optical satellite imagery to efficiently and semi-automatically detect detailed vector habitats. A WorldView2 satellite image capable of <5m resolution was acquired in the spring of 2013 for the area around Bruges, Belgium, a region where dairy farms suffer from liver fluke infections transmitted by freshwater snails. The vector thrives in small water bodies (SWBs), such as ponds, ditches and other humid areas consisting of open water, aquatic vegetation and/or inundated grass. These water bodies can be as small as a few m2 and are most often not present on existing land cover maps because of their small size. We present a classification procedure based on object-based image analysis (OBIA) that proved valuable to detect SWBs at a fine scale in an operational and semi-automated way. The classification results were compared to field and other reference data such as existing broad-scale maps and expert knowledge. Overall, the SWB detection accuracy reached up to 87%. The resulting fine-scale SWB map can be used as input for spatial distribution modelling of the liver fluke snail vector to enable development of improved infection risk mapping and management advice adapted to specific, local farm situations.

Abundance modelling of invasive and indigenous Culicoides species in Spain.

In this paper we present a novel methodology applied in Spain to model spatial abundance patterns of potential vectors of disease at a medium spatial resolution of 5 x 5 km using a countrywide database with abundance data for five Culicoides species, random regression Forest modelling and a spatial dataset of ground measured and remotely sensed eco-climatic and environmental predictor variables. First the probability of occurrence was computed. In a second step a direct regression between the probability of occurrence and trap abundance was established to verify the linearity of the relationship. Finally the probability of occurrence was used in combination with the set of predictor variables to model abundance. In each case the variable importance of the predictors was used to biologically interpret results and to compare both model outputs, and model performance was assessed using four different accuracy measures. Results are shown for C. imicola, C. newsteadii, C. pulicaris group, C. punctatus and C. obsoletus group. In each case the probability of occurrence is a good predictor of abundance at the used spatial resolution of 5 x 5 km. In addition, the C. imicola and C. obsoletus group are highly driven by summer rainfall. The spatial pattern is inverse between the two species, indicating that the lower and upper thresholds are different. C. pulicaris group is mainly driven by temperature. The patterns for C. newsteadii and C. punctatus are less clear. It is concluded that the proposed methodology can be used as an input to transmission-infection-recovery (TIR) models and R0 models. The methodology will become available to the general public as part of the VECMAP™ software.

Factors affecting Bluetongue serotype 8 spread in Northern Europe in 2006: the geographical epidemiology.

In 2006, Bluetongue serotype 8 was notified for the first time in north-western Europe, more specifically in Belgium, the Netherlands, Luxemburg, Germany and France. The disease spread very rapidly, affecting mainly cattle and sheep farms. In this paper, we examined risk factors affecting the spatial incidence of reported Bluetongue events during the first outbreak in 2006. Previous studies suggested that the Bluetongue incidence was enhanced by environmental factors, such as temperature and wind speed and direction, as well as by human interventions, such as the transport of animals. In contrast to the previous studies, which were based on univariable analyses, a multivariable epidemiological analysis describing the spatial relationship between Bluetongue incidence and possible risk factors is proposed in this paper. This disentangles the complex interplay between different risk factors. Our model shows that wind is the most important factor affecting the incidence of the disease. In addition, areas with high precipitation are slightly more sensitive to the spread of the infection via the wind. Another important risk factor is the land cover; high-risk areas for infection being characterized by a fragmentation of the land cover, especially the combination of forests and urban areas. Precipitation and temperature are also significant risk factors. High precipitation in areas with a large coverage of forests and/or pasture increases the risk whereas high temperature increases the risk considerably in municipalities covered mainly with pasture. Local spread via the vector is strongest in areas with a large coverage of forests and smallest in highly urbanized areas. Finally, the transport of animals from infected areas is a risk factor.

Improving the accuracy of livestock distribution estimates through spatial interpolation.

Animal distribution maps serve many purposes such as estimating transmission risk of zoonotic pathogens to both animals and humans. The reliability and usability of such maps is highly dependent on the quality of the input data. However, decisions on how to perform livestock surveys are often based on previous work without considering possible consequences. A better understanding of the impact of using different sample designs and processing steps on the accuracy of livestock distribution estimates was acquired through iterative experiments using detailed survey. The importance of sample size, sample design and aggregation is demonstrated and spatial interpolation is presented as a potential way to improve cattle number estimates. As expected, results show that an increasing sample size increased the precision of cattle number estimates but these improvements were mainly seen when the initial sample size was relatively low (e.g. a median relative error decrease of 0.04% per sampled parish for sample sizes below 500 parishes). For higher sample sizes, the added value of further increasing the number of samples declined rapidly (e.g. a median relative error decrease of 0.01% per sampled parish for sample sizes above 500 parishes. When a two-stage stratified sample design was applied to yield more evenly distributed samples, accuracy levels were higher for low sample densities and stabilised at lower sample sizes compared to one-stage stratified sampling. Aggregating the resulting cattle number estimates yielded significantly more accurate results because of averaging under- and over-estimates (e.g. when aggregating cattle number estimates from subcounty to district level, P <0.009 based on a sample of 2,077 parishes using one-stage stratified samples). During aggregation, area-weighted mean values were assigned to higher administrative unit levels. However, when this step is preceded by a spatial interpolation to fill in missing values in non-sampled areas, accuracy is improved remarkably. This counts especially for low sample sizes and spatially even distributed samples (e.g. P <0.001 for a sample of 170 parishes using one-stage stratified sampling and aggregation on district level). Whether the same observations apply on a lower spatial scale should be further investigated.