Coordinator: Ignasi Bartomeus (EBD-CSIC)
In recent years we’ve made a huge advance on understanding the effect of managed and wild pollinators on crop yield. However, our ability to predict visitation rates and crop yield are still limited due to the huge variation observed among crops, years and environments. To advance in this direction we want to (i) compile a dynamic and open database on crop pollination data published as a data paper with all data contributors as co-authors and (ii) use this database to compare the predictive ability of data-driven, statistical and mechanistic models on crop pollination.
We want to reach all researchers working on the pollination of different crops to create a dynamic crop pollination database. Basically, we are looking for studies measuring pollinators within the crop (visitation rates or abundances) for at least five fields and ideally a measure of crop production (tonnes per hectare, fruit set, …). Here you can download a template with the requested data. However, if you have data in another format and no time to transform it to the template format, contact us and we can find a solution or we can directly accept raw data. Data should be sent to Alfonso Allen-Perkins (firstname.lastname@example.org) and queries to Ignasi Bartomeus (email@example.com). We will not publish anything without your explicit consent and you can always draw back your data in a later stage if you wish to do so.
Open data papers allow us to gather global standardized data on a particular topic, thus constituting a huge resource for the scientific community. These papers are highly cited products, and the outcomes of the analysis that can be potentially done with them, especially when merged with other types of data, are unpredictable. This approach also represents a great way to give credit to everyone involved in the collection of data, as all data collectors are co-authors of the resulting data paper. In addition, we will encourage the citation of each individual study when only those are used. Examples of such efforts are the predicts database, TRY or the meta-community database. We think that by building a worldwide crop pollination database (which will keep growing in the future years) we can stimulate the advance of science and give proper credit to all data holders. We know data collection is a huge effort, and before publishing the data, we will share with all co-authors the full database at least 6 months in advance. We do not expect to publish the data paper until the end of 2021 to allow time to get as much data as possible.
As mentioned above, we plan to use this data to compare the predictive ability of different modeling techniques. Prediction is the next frontier in ecology and if successful, our work can have a great impact. Given this will be the first use of the database, we would like to invite interested data holders that also want to get involved in the manuscript to co-author this first global paper. We expect the paper to be published in 2022.
We hope to have you on board. Do not hesitate to ask any relevant question.
We already compiled data on more than 160 study systems worldwide, comprising > 2000 fields from 38 different crop species. You can see there are some geographical gaps, share your data!
Do you accept published datasets? Yes, in fact most datasets are already published and we will link your dataset to the original publication
Do you accept data collected within cereal fields or greenhouses? Yes, while cereals do not depend on pollinators, this data may be interesting for questions about diversity in agricultural landscapes. On the other hand, greenhouse data may not be relevant for diversity questions, but can help relating visitation rates to yield.
I have data for several plots within each field, do you also want this resolution? Most datasets contain agreggated measures per field. Given the different experimental dessigns used by diferent researchers, aggregated data is the easiest to armonize across datasets and the preferred format.
Do you want pan trap data? We updated the template to capture the sampling method. While pollinator abundance or visitation rates are more direct methods. Pan trap data or unestrcutured sweep netting data are also interesting.
I have a different yield measure than the prefferred one, what can I do? We understand that each study was dessigned different and different crop yields are measured in different units. While seed set (%) and weigth per hectare are (tones/ha) are quite common, if you have other measures, we added a column to specify your measure and allow up to two units to be reported by treatment (open pollination, pollinator exclusion, and pollen suplement).
1- A dynamic and open global database on crop pollination. In preparation. Expected publication date: 2021.
2- Modelling pollination services: A model performance comparision. In preparation.