Identify vulnerability model for water stress (agro-climate model) and accompanying hazard data set

[joemoorhouse]

@aliebadi22, I think you were looking at some Copernicus data for drought (or possibly water stress?)? Creating this issue as a place to share your refs and more discussion.

[aliebadi22]

CDD is a drought monitoring and drought damage indicator that counts the maximum number of Consecutive Dry Days, with daily precipitation amount of less than 1 mm. i.e., drought spell. This climate index is a measure of low precipitation, with high values corresponding to long periods of low precipitation and potentially drought-favouring conditions. An increase of this index with time means that the chance of drought conditions will increase.

The free access to to the data is provided by Climate Data Store which is a free data service and toolbox. The mentioned indicators are agroclimatic indicators from 1951 to 2099 derived from climate projections. These data are provided in .nc formats which is usually used in engineering designing and manufacturing. You can find the data requests to download in the following link. Also, an API and Toolbox for download is provided.

Agroclimatic indicators from 1951 to 2099 derived from climate projections (copernicus.eu)

A total of 26 indicators are provided, covering the global land area at the spatial resolution of 0.5°x0.5° lat-lon grid (Almost 50 km2). As part of the C3S Global Agriculture SIS, the agroclimatic indicators are generated to represent features of the climate that are used to characterise plant-climate interactions. Agroclimatic indicators are useful in conveying climate variability and change in the terms that are meaningful to agriculture.

There are two versions of data v1.0 and v1.1 with little difference. The frequency of data for the chosen indicators are seasonal. The available models are: GFDL-ESM2M, HadGEM2-ES, IPSL-CM5A-L, MIROC-ESM-CHEM, and NorESM1-M.

To connect a location to a CDD value, one can fine the closest point to the grid location on the dataset that is possible using Haversine formula. Also, one can do it quickly using geopandas package of python. The implementation of the Haversine formula is explained in this link:

https://nathanrooy.github.io/posts/2016-09-07/haversine-with-python/

The percentage change from the baseline scenario is one way to study the CDD. Moreover, studying the evolution over the years or the change of value between the scenarios might be useful.

[joemoorhouse]

As discussed with @floriangallo and @Obsidjan, probably we want to start with a regional as opposed to global model

A couple of interesting references:

Assessment on Agricultural Drought Vulnerability and Spatial Heterogeneity Study in China - PMC (nih.gov)

Home | AWA- AgriAdapt platform for adapting farms to climate change

[floriangallo]

http://web.stanford.edu/~mburke/papers/Burke_Emerick_2015.pdf

[joemoorhouse]

NHESS - Global-scale drought risk assessment for agricultural systems (copernicus.org)

Shall we align (long-term) climate models with the near-real-time?
https://earthdaily.com/

[joemoorhouse]

From @mariembouchaala:
https://www.institutdesactuaires.com/docs/mem/daeb4e950247915317c656f14a4e1773.pdf
[Interesting as includes the financial model]

[joemoorhouse]

Also look at IPCC working group 2: good source of references
https://www.ipcc.ch/report/sixth-assessment-report-working-group-ii/

[floriangallo]

https://www.ipcc.ch/report/ar6/wg2/downloads/report/IPCC_AR6_WGII_Chapter05.pdf

[Obsidjan]

Some regional analysis examples below: (can add them to the lit review, but we could try to implement one of these. We essentially need to generate vulnerability criteria for whatever crop we are looking at, then combine that criteria into an integrated model that can plug in the known values of agri assets (most investors aren't going to be invested in one crop type or region only).

• Using the EPIC model, Hao Guo et al. constructed a database suitable for large-scale risk assessment, fitted the loss rate-drought index-environmental index (L-D-E) vulnerability surface with irrigation scenarios, and quantitatively evaluated the global corn drought risk based on the optimal vulnerability surface of corn [1,2,3].
• Based on the definition of vulnerability by the Intergovernmental Panel on Climate Change (IPCC), Yan Li et al. used the entropy method to evaluate the drought vulnerability of maize in Northwest Liaoning.
• Philip et al. used rainfall and yield data as well as proxy indicators of adaptability to combine various aspects of crop drought vulnerability with socio-economic indicators to evaluate the exposure, adaptability, sensitivity, and vulnerability of 10 regions in Ghana.4
• Hahn et al. applied the livelihood assessment method to the vulnerability assessment of climate change in Mozambique, Eastern Africa, and calculated the vulnerability index defined by the intergovernmental panel on climate change (IPCC) 5,6,7].
• Khaled Hazaymeh et al. developed a remote sensing-based agricultural drought indicator (ADI) at a 30-meter spatial resolution and 8-day temporal resolution, and evaluated its performance in Jordan’s semi-arid region 8].
• Yin Gao et al. establish a multi-index method to evaluate drought vulnerability of maize in different growth periods in Central and Western Jilin Province. Based on the existing meteorological data, the yield of corn in the counties and cities where the meteorological stations are located and the proxy indexes of adaptability, the drought disturbance degree, sensitivity, self-recovery ability, and environmental adaptability of the central and western Jilin Province are evaluated. By simplifying and selecting indicators, this vulnerability assessment model is applied to regions lacking more detailed data. 9

[joemoorhouse]

Hi @reilly335, you had a good suggestion: that for agriculture vulnerability models we can try to align with Earth Daily Agro https://earthdailyagro.com/. That is, can we have a modelling approach consistent with those used in near real-time monitoring, even if we are focussed on climate-change scenarios. Do you have any contacts who would be happy to discuss modelling approach with us?

[joemoorhouse]

Hi @Obsidjan, @floriangallo,
Although implementing a model like that of Yin Gao et al. might be a good way in, do we rather/also want to use projected crop yields directly from ISIMIP outputs (although it looks like latest data subject to embargo period)?
https://www.isimip.org/outputdata/