Assessment of map quality

A common method to assess the quality of a course scale data-set is to compare it with independent information at selected locations on a more detailed scale in order to draw conclusions with respect to the map quality at these locations and extrapolate it to the general map quality. Here, however, all available data on irrigated areas at appropriate scales were used to compile the map itself and therefore could not be used for a quality assessment. Apart form this, it is very difficult to compare the quantitative information as presented on the map (expressed as area equipped for irrigation in a 5-minute cell in ha) with point information collected for specific positions by ground-truthing (irrigated or rainfed).

To assess the quality of the Global Map of Irrigation Areas, two indicators were developed that take into account the geospatial information density of the base data:

• Indicator A (IND_A), represents the density of the used subnational irrigation statistics;
• Indicator B (IND_B), represents the density of the available geospatial records on position and extent of irrigated areas;

The combination of IND_A and IND_B is used to describe the overall map quality per country.

The density of subnational irrigation statistics can be obtained by calculating the arithmetic mean of the size of the subnational units. However, there are some countries where irrigation is concentrated in some small subnational units while in other very large subnational units of the same country there is no or very little irrigation. To avoid that large subnational units without significant irrigation have a negative impact on indicator A, the size of each subnational statistical unit is weighted by the irrigation density in the subnational unit relative to the irrigation density in the entire country:

with:

where IND_A_country is the average weighted size of the subnational units in the specific country (ha), areacountry is the surface area of the country (ha), irridensadm is the irrigation density in subnational unit adm (-), irridenscountry is the irrigation density in region the country (-), n is the number of subnational units in the country, irareaadm is the irrigated area in subnational unit adm (ha) and areaadm is the surface area in subnational unit adm (ha).

IND_A equals the average size of all subnational units in a country if the irrigation density is the same all over the country the country. If all irrigated area is concentrated in only one subnational unit, IND_A is equal to the size of this subnational unit. IND_A is lower than the average size of the subnational units if the irrigation density is higher in small subnational units than in the larger subnational units. Lower values of IND_A indicate a better map quality.

The second indicator (IND_B) was developed to give an estimate on the density of geospatial information used to assign irrigated area to specific cells within the sub-national units. IND_B was computed as the fraction of irrigated area that could be assigned to specific grid cells by using geospatial records on the position and extent of known irrigation projects. Higher values of IND_B indicate a better map quality.

Both indicators were assigned a country mark according to the classification in the following table:

A mark for the overall quality was given assuming that the types of information that are reflected by the two indicators can replace each other. The mark for the overall map quality was set to the better of the two marks given according to IND_A and IND_B. If, for example, the location and extent of almost all irrigation projects in a country is known then the overall quality of the map should be excellent independently from the mark given according to the weighted size of subnational units. On the other hand, if the size of the subnational statistical units is very small (in an extreme case smaller than the map resolution of 5 arc minutes), the overall quality of the map should also be excellent even if there are no geospatial records on the position of irrigation schemes within the sub-national units available. Finally the mark for the overall map quality was downgraded when there were doubts regarding the reliability of the information used for a specific country. Eventually all countries got a mark from 1 to 4 representing "Very Good" to "Poor". In version 4 of the Global Map of Irrigation Areas none of the countries was classified as either "Excellent" or "Very Poor". The quality marks for each country are listed in the table below.

Marks for the overall mapping quality in world regions or at global scale were computed by combining the marks for the overall quality of the map at country level and the irrigated area in the corresponding countries as:

where m_reg is the overall quality of irrigation map in the whole region, irarea_{v_good}, irarea_good, irarea_fair and irarea_poor represent the irrigated area of all countries in the region with very good, good, fair or poor map quality (ha) and irarea_reg is the irrigated area in the whole region (ha). The table below shows the map quality of different regions:

A detailed explanation of the quality assessment of the Global Map of Irrigation Areas is described in the paper Development and validation of the global map of irrigation areas published in Hydrology and Earth System Sciences. It should be noted that the analyses for this paper were done on map version 3.

Your access to AQUASTAT and use of any of its information or data is subject to the terms and conditions laid down in the User Agreement.