Downscaling livestock census data using multivariate predictive models: sensitivity to modifiable areal unit problem

Abstract

The analysis of census data aggregated by administrative units introduces a statistical bias known as the modifiable areal unit problem (MAUP). Previous researches have mostly assessed the effect of MAUP on upscaling models. The present study contributes to clarify the effects of MAUP on the downscaling methodologies, highlighting how a priori scales and shapes choices could influence the results. We aggregated chicken and duck fine-resolution census in Thailand, using three administrative census levels in regular and irregular shapes. We then disaggregated the data within the Gridded Livestock of the World analytical framework, sampling predictors in two different ways. A sensitivity analysis on Pearson's r correlation statistics and RMSE were carried out to understand how size and shapes of the response variables affect the goodness-of-fit and downscaling performances. We showed that scale, rather than shapes and sampling methods, affected downscaling precision, suggesting that training the model using the finest administrative level available is preferable. Moreover, datasets showing non-homogeneus distribution but instead spatial clustering seemed less affected by MAUP, yielding higher Pearson's r values and lower RMSE compared to a more spatially homogenous dataset. Implementing aggregation sensitivity analysis in spatial studies could help to interpret complex results and disseminate robust products.

dynamAedes: a unified modelling framework for invasive Aedes mosquitoes

“…Our modelling approach focused on the "species fundamental thermal niche" concept (sensu Hutchinson [118]) by considering temperature as the main driver of population dynamics. In light of this, the choice of temperature datasets is crucial [119]; for example, pixel size may influence model outcome because of the aggregating effect of the Modifiable Unit Areal Problem [120][121][122]. Climatic reanalysis and global/regional circulation models are reliable data sources with high temporal resolution for present climatic conditions and robust future projections, but they have coarse spatial resolution that may underestimate the microclimate and its effects on species biology [15,123].…”

Assessment of Population‐Level Disadvantage Indices to Inform Equitable Health Policy

“…Returning to the MAUP and the issue of ecological fallacy, users should carefully consider how the level of geographic analysis will impact interpretation of results when using disadvantage indices. Resolving biases that result from the aggregation of spatial data requires critically appraising geographic scope and using the smallest groupings, where possible, to examine variables for mapping 71 . Before making allocation decisions or implementing interventions, policymakers should analyze local clustering and context to ensure appropriate interpretation of neighborhood‐level disadvantage.…”

“…Resolving biases that result from the aggregation of spatial data requires critically appraising geographic scope and using the smallest groupings, where possible, to examine variables for mapping. 71 Users should also consider that geographic boundary definitions might not be sensitive to demographic changes due to displacement or sociopolitical processes like gentrification or gerrymandering. Further, as noted in other research, many patterns of social risk overlap with historic redlining practices and racial housing segregation.…”

dynamAedes: a unified modelling framework for invasiveAedesmosquitoes

“…Secondly, we recognise that pixel size may influence model outcome because of the aggregating effect of the Modifiable Unit Areal Problem (Jelinski and Wu, 1996; Da Re et al, 2020). While the consequences of this artifact are well known in SDMs applications, they are rarely mentioned or addressed (but see Peterson, 2014).…”