Our Organisation Search
Quick Links
Toggle: Topics

GIS Monthly Maps

Map of the Month July

Cartographer: Dr Stuart Green

To the Map: An indicative history of green cover since the 1980's

 

The management history of any parcel of land is an important piece of information. Knowing the history can help us understand issues around greenhouse gas accounting, bio-diversity, soil health and many more.

This month we present an interim output from a study looking to see how often fields are bare of vegetation, due to harvesting or re-seeding. Using the Teagasc archive of satellite imagery going back to the 1980's we can use vegetation indices to give an indication of whether vegetation is present or not.

Because of cloud cover each field is only imaged perhaps 4 times a year, so the analysis can only give an indication of how often the parcel has had low vegetation cover.

The archive used had 317 Images over 35 years up to 2019, each parcel had on average 210 observations. Each image was used to populate the Prime2 vegetation polygons (from 2017), with the average NDVI* value for each pixel in the image. The archive is quality controlled for atmospheric and cloud issues, so only good quality pixels are used. A cut-off value for bare fields was established, referenced against total image means. The number of time the reference value was met was counted for each field and this is the number of times the field has appeared bare in our satellite record.

The dark green fields in the map have been covered with vegetation almost permanently in this record, with fewer than 2 low observations. They would potentially be sites of botanical interest as they may less intensely managed than other grasslands in the area. The pink fields have been apparently bare in more than 10% of the observations. These fields are mostly tillage but some are intensely managed grassland. A few of the pink areas around Limerick city are due building in the city recently and some may be capturing a record of flood events. The rest of the fields have a range of counts of being bare- we chose to only highlight the two extremes in the map.

This is a partial, interim output- the full data set, with complete time-signatures of bare soil should be available at the end of the year.

*NDVI or Normalised Difference Vegetation Index, is a long established method to monitor vegetation cover from satellites. It relies on the fact that vegetation deeply absorbs red light but strongly reflects the very near infra-red (a wavelength slightly longer than visible red light). The ratio of the two values is the vegetation index and the higher the number, up to 1, the greater the amount of vegetation.

https://en.wikipedia.org/wiki/Normalized_difference_vegetation_index

 

Map of the Month June

Cartographer: Dr Jesko Zimmermann

To the map: Drought stress following low rainfall levels in late spring 2020

Drought has a severe impact on Irish agriculture as Irish farming systems are reliant upon regular rainfall. The dry summer of 2018 (in combination with a cold spring) for example has led to a severe shortage in animal feed due to reduced grass growth and the depletion of fodder stocks.

It is therefore not surprising that the current dry spell is causing concern within the farming community. Since early March 2020 parts of the country have seen very little rainfall with rainfall rates up to 90 % lower than the 1980 to 2010 reference period. This is impacting grass growth as well as other farming enterprises such as horticulture. As in 2018, the impact of the drought is not equally distributed across Ireland with many factors impacting the severity of the drought. These include the soil type, where heavy soils retain water much longer under dry conditions, while lighter sandy soils will dry quicker. Furthermore, weather patterns are not equal with some parts of the country (such as the south west) showing more normal rainfall levels. Finally, management also determines how much an area is impacted by low rainfall levels. Irrigation will offset the impacts, while drainage may worsen the drought effects.

In this map we use a proxy for drought stress, specifically the Normalised Difference Moisture Index (NDMI), to highlight the spatial variability in severity of the drought. The NDMI can be derived from multispectral satellites (in this case data the NASA MODIS sensor) and uses the reflectance from the near infrared and the short-wave infrared bands to estimate plant moisture content. A lower value indicates a lower plant moisture content and therefore potential drought stress. For easier interpretation, we related the average NDMI from May 2020 to the average May NDMI for the previous ten years (2009 to 2019). A negative value means the current NDMI was lower (less water in plants), and a positive values mean more water in plants.

Overall the NDMI was marginally lower in 2020. The spatial distribution, however, is of more interest showing a large area in the east and north-east, as well as several pockets in the west with drier conditions than usual. The analysis also shows areas with higher than average NDMI which are mainly situated in the upland areas and generally the south-west. The general south-west/north east gradient reflects the overall rainfall patterns in the last months. The pockets of drier and wetter conditions are likely linked to soil conditions (wetter areas occur especially in uplands and peaty areas, while drier areas in the west such as in Co. Limerick coincide with well drained soils.)

An animation of the underlying data can be found on the Earth Observation blog (https://earthobservation.wordpress.com/2020/06/11/ndmi-in-the-past-ten-years/). It shows the NDMI averages for each May from 2009 to 2020. 

 

Map of the Month May

Cartographer: Dr Stuart Green

To the map: Distribution of Cultivated Peats

In this months map we look at the distribution of cultivated peat soils. Whilst peat-lands which include raised bogs and blanket bogs and fens,  cover upwards of a fifth of the country, their use can be limited because they are generally very wet in the un-drained state. The extraction of peat for fuel, either mechanically or by hand is a principal use of these soils but this is drawing to a close as the climate impacts of draining and burning peats are now clear. In their natural state peats are natural sinks of carbon and are the largest reservoir of stored carbon in the Irish landscape. The remaining bogs are seen as vital parts of Irish biodiversity and so are no longer planted with forestry which has traditionally been a significant land use of peat soils. The blanket bogs cover much of the commonage areas in the country and are thus used for (and indeed maintained by) rough grazing of sheep.

However not all peat soils are bogs, a significant portion of peat soils have been used and adapted over the decades through drainage and fertilization for grass (and to a much smaller extent cereal) production. These grasslands over peats are no longer bogs in an environmental sense but they are primarily organic soils and would revert back to a natural vegetative state were it not for continuous management through drainage, grazing and fertilizing/liming.

Irish research has shown that this sort of management of organic soils, especially drainage, turns these natural stores of carbon into emitters of carbon so in order to get the correct accounts of emissions and sequestrations of land use it’s vital that we know where these cultivated organic peat soils are.

To map the distribution we used the peat soils in Teagasc Indicative Soil map, along with field boundaries mapped in the OSI National Digital Map database and parcel information in LPIS. This allowed us to identify fields, currently farmed, that overlay peat soils. The image here gives an example of 3 bare fields with a mix of soils with the darker peat soils to the right.

We excluded non-farmed bogs, commonage areas, forested areas and worked bogs (peats extracted for fuel).

To get a sense of how heavily managed and thus how productive the fields are we looked at satellite imagery for 2018 (a drought year, but one in which peat soils and poorly drained mineral soils performed well because of their moisture content). Using a method of analysis called vegetation indexes we can use the satellites to score fields between 0 and 1, with a higher score meaning more biomass production. We banded the fields into 3 bins; Low, Medium and High based on these indexes to give a crude indication of the level of production.

Whilst the analysis is done at field scale, the map is presented at 1km square scale- each 1km cell is coded for the percentage of cultivated peats in the cell and whether they are mainly high, medium or low production. We have omitted cells with less than 30% cultivated peats to make the distribution clearer and not to give a false impression of the true areal extent of cultivated peats.

This analysis suggests approximately 6% of the country or 420,000Ha is made up of cultivated peats across a wide range of farming intensities (though predominately low intensity farming).

Map of the Month April

Cartographer: Dr David Meredith

To the map: Covid-19: Distribution of Populations with Higher Risks

 

 Background

With the ongoing success of measures to reduce the spread of the Covid-19 virus, attention is now being given to re-opening businesses, schools and social services. At the same time, the World Health Organisation has highlighted that this virus will be with us for some time. Studies of previous pandemics suggest that we may face a number of ‘waves’ of infection. It is apparent from current trends that older age groups, people with poor health or who are already ill and those living in some types of communal establishments are particularly susceptible to Covid-19. This highlights one of the key challenges of managing future waves of infection, i.e. how to limit the spread within the overall population and, particularly, to these vulnerable populations. Whilst it is to be expected that strict controls in place in some communal residential settings will be effective in helping to protect their residents, the challenge of restricting spread in the general community will persist for some time to come.

 

Map

We show the result of an analysis that assessed the proportion of the population of each Small Area that was over 64 years of age, unable to work due to sickness or disability, reported their health status to be less than ‘Good’ and  persons living in communal accommodation on the night of the Census (2016). For each indicator, we ranked Small Areas i.e. from 1 (lowest indicator value) to 18,641 (highest). We then combined these ranks to give an overall score. The classification of Small Areas into the different categories is based on a decile distribution, i.e. the 10% of Small Areas with the lowest scores are shown in Yellow whilst the 90%+ category shows the 10% of areas with the highest scores.

The map points to the concentration of populations that are particularly vulnerable to Covid-19 and highlights those areas that could be severely impacted by an infection if it were to spread within their populations. The broad pattern depicted on the map highlights rural areas outside of the major commuter zones and towns or inner city areas as recording higher vulnerability scores. This reflects the distribution of the population over 65 years of age, and concentrations of those unable to work and those with poor health.

 

Data

The data used in this analysis are produced by the Central Statistics Office from the Census of Population (2016). The statistical data are drawn from the Small Area Population Statistics and are available here: https://www.cso.ie/en/media/csoie/census/census2016/census2016boundaryfiles/SAPS2016_SA2017.csv.The spatial data are also available through the CSO here: https://data.gov.ie/dataset/small-areas-generalised-20m-osi-national-statistical-boundaries-2015.

There are 18,641 Small Areas and the CSO publishes summary data from the Census of Population for each of these. These indicators used in this analysis were selected as they, very broadly, reflect the profile of particularly vulnerable groups to Covid-19. No attempt was made to weight the relative importance of the indicators. A key limitation of this analysis is that it does not attempt to estimate the risk of infection by weighting the vulnerability score for each area by taking into consideration the population density of each area or the level of interaction between areas. Analysis is on-going to take these issues into consideration.

Note:

This analysis is an extension of Teagasc’ Rural Health Research programme undertaken by Teagasc with additional support from the Department of Agriculture, Food and the Marine. The focus of this programme is on supporting improved farm safety and the wellbeing of farmers and farm families.  

Map of the Month March

Cartographers: Dr Stuart Green

March 2020 Map of the Month

As we get used to the idea of social distancing this month’s map reminds us that in many parts of the country physical distance is an issue. The Map is intended to show rural isolation, using as an indicator a 2km circle around every home in the country (our new personal geographic boundary). In this 2km boundary we counted the number of other houses (and did this for every house in the country).

A 2km radius circle is quite a bit area- 1200Ha, and within cities there are 1000's of residences in this area but in parts of rural Ireland the number of dwellings can get very small and the map highlights those area where each household has fewer than 20 neighbours. In a city or town, even when staying at home, there are neighbours next door, people walking past the window- someone to talk too over the garden fence. In these remote areas, being at home means real isolation.

These areas are remote, not only from neighbouring houses but all kinds of services and therefore pose particular issues adhering to the new social distancing/cocooning rules.

 

Map of the Month February

Cartographers: Dr Jesko Zimmermann & Dr Rob O'Hara

Arable farming then and now

The map shows the centre point of Irish townlands with a reference to arable land in their names, and compares them to the current distribution of arable land (defined as the land at least once used for arable agriculture from 2000 to 2016). The names of places can provide valuable insights into local history and in this case can provide a proxy for past agricultural practices.

In this case we identified a number of Irish words that indicate arable practices and filtered the Irish language townland names for those containing said words. We used P. W. Joyce's book Irish Names of Places as a guideline for such references. In this analysis we limited the search to the Irish names of townlands, as the Anglicisation often obscures the Irish origins, with single anglicised forms originating from multiple words. We made an exception for Northern Ireland, where no Irish townland names were easily available. To avoid false positives we only included unambiguous anglicised terms (e.g. Cappa or Mullen in varying versions).

The current extent of cropland was derived from the Land Parcel Identification System and is displayed as the share of land under crop for at least one year between 200 and 2016 within a 2.5 x 2.5 km cell.

The map shows a number of interesting patterns. The majority of townlands names referring to arable land occur west of Ireland, while occurrence in the current extent of cropland is relatively low. This does not indicate a shift in the extent of cropland but rather a different naming culture in the east of Ireland which was under Norman control. The Irish 'gort' (a tilled field) being the most common reference occurring evenly in Connacht and Munster. It should, however, be note that the term ‘gort’ is subject to some ambiguity when looking at different sources. While P.W. Joyce translates it as ‘tilled field’, the townland register (https://www.logainm.ie/) translates ‘gort’ simply as ‘field’. ‘Ceapach’ (tilled ground) is less common but shows a similar distribution across the west. ‘Tamhnach’ occurs mainly in Connacht. Other references are very rare, except for ‘muileann’ which refers to a mill which is the most common reference in Northern Ireland. This is likely the result of excluding references with ambiguous anglicisations.

Map of the Month January

Cartographers: Dr Stuart Green

Landuse Change in Co. Roscommon 1995-2017

The first map of 2020 provides a snapshot of changes in landuse in County Roscommon

As part of the EPA funded SOLUM project we have been looking at landuse change across the country. One output is a map of landuse change in County Roscommon. We recently reviewed original sample points from a 1995 study, both a regular grid sample and other points focused on individual
landcover classes (total of 1024 points.) looking for changes in grassland management, as well as collecting data on points that became forested since 1995. We also found one site that has become a tilled cropland since 1995, and unusually a new waterway - a marina built on the Shannon.

In total 23% of the gridded samples showed some change. In the south of the county there was little change. Across the middle of the county, south of the drumlin belts there is evidence of grasslands being more intensely managed. Across the north of the county, with peat soils being common, there is a larger percentage of new forestry and woodland. Whilst some grassland sites in this region are showing greater degrees of management it is more common for changing grassland sites to show less management now than in 1995 (whilst remembering 77% of sites haven't changed).

About Map of the Month

In addition to undertaking geographical analyses and producing maps for research projects, the spatial analysis lab responds to ad hoc requests for contributions.  The latter may be for in-house purposes or to inform policy submissions. While dissemination is a key objective of research projects, maps produced in response to such requests rarely get a wider audience. We’ve decided that we’ll take the most interesting map we have produced in each month and to present it here to hopefully find a wider audience and promote discussion and debate on both the contribution of spatial analysis to Irish agriculture and food and on the specific maps produced.

Whilst this map can be shared please check with us before reproducing it in a publication. Many of the data sets we use are under licence with conditions attached.

For general enquiries contact Stuart Green or the author above for information on this month’s map.