Simulating cropping sequences using earth observation data

Exploring the effects of land management change on productivity, carbon and nutrient balance: Application of an Ensemble Modelling Approach to the upper River Taw observatory, UK

“…Lambs were assumed to be finished at 40 and 38 kg in the RLR and moorland respectively and beef were assumed to have a finishing weight of 517.65 kg ( Nix, 2020 ). The RLM contains a novel crop rotations generator which, based on the crops grown in a given area (in this case the Taw catchment between 2016 and 2019, see Tables S4 and S5) and certain agronomic rules, stochastically generates realistic crop sequences for that area ( Sharp et al, 2021 ). In this way realistic rotations of crop production for a particular area can be simulated.…”

A new approach to characterising and predicting crop rotations using national-scale annual crop maps

“…This is likely due to economic pressures to grow higher-profit crops in shorter rotations, despite lower yields (Hegewald et al, 2018). If we restricted outcomes using agronomic expertise (as Sharp et al, 2021), we might not have seen this or other patterns emerging from the data. In comparison to our work, we found existing studies characterising or predicting crop rotations have been more limited in at least one way: coarser spatial resolution (e.g.…”

“…There has been previous work investigating crop sequences where the main point of interest is in defining the likelihood of which crop is likely to follow another in GB (Sharp et al, 2021). However, existing approaches do not take into account further history of the site or field so that the probability of crop classification at time t is conditionally independent of the crop classification at time t-2, t-3, …,t-n, given the crop at time t-1.…”

“…However, crop rotation data are generally limited in their spatial coverage, resolution, or are not spatially explicit. Existing approaches to crop prediction modelling have often focused on predicting a single crop type depending on the crop in the previous year or two (Castellazzi et al, 2008;Detlefsen and Jensen, 2007), via a range of methods including transition probability matrices (TPMs; Castellazzi et al, 2008;Detlefsen and Jensen, 2007;Kim et al, 2020;Mignolet et al, 2007;Osman et al, 2015;Zhang et al, 2019;Sharp et al, 2021), rulebased methods (Bachinger and Zander, 2007;Detlefsen and Jensen, 2007;Dogliotti et al, 2003;Schönhart et al, 2011), machine learning (Osman et al, 2015;Zhang et al, 2019), and spatial learning (Yaramasu et al, 2020). However, few studies have tested the relative accuracy of different spatial approaches to refine predictions, or assessed patterns at a national scale.…”

Mapping the Complex Crop Rotation Systems in Southern China Considering Cropping Intensity, Crop Diversity, and Their Seasonal Dynamics

“…And in the second step, the generated crop sequences are spatially allocated according to specific rules to obtain the spatial patterns [17], [18]. However, modelling researches often lack actual data on crop sequences and thus need to rely on the analyst defining realistic crop sequences, which may not account for the large variations in farming practices across a region [19].…”