New Brunswick County Map with Cities, Rivers and Population Density

Heading to the field? Install this site from your browser menu and this tool works with no signal. See how

How to use this tool

Pick a map type along the top and the map redraws. Divisions colours the fifteen counties so no two neighbours share a fill, Density shades them by people per square kilometre, Cities sizes each town by population, Rivers draws the main watercourses, Satellite swaps the flat colours for real satellite imagery of the province, and Blank outline gives you black on white for printing. Click any county, or any row in the table, and its figures appear in the Selection panel.

The layer switches on the right are independent of the map type. That is the part people miss. Choosing a type sets a sensible starting combination, but every switch still works afterwards, so a blank outline with rivers turned on and boundaries turned off is two clicks away and it is the version most people actually want to print. Move the pointer across the map and the readout at the bottom left gives you latitude and longitude in degrees, minutes and seconds.

Where the data comes from

Two agencies, joined on one key. The county polygons are the GeoNB county file published by Service New Brunswick. The population and land area figures are Statistics Canada, 2021 Census of Population. They meet on the census division identifier, CDUID 1301 through 1315, which runs Saint John, Charlotte, Sunbury, Queens, Kings, Albert, Westmorland, Kent, Northumberland, York, Carleton, Victoria, Madawaska, Restigouche, Gloucester.

Geometry and statistics sit in separate files rather than one merged layer. That sounds like extra work and it saves work later: adding median income or dwelling counts becomes a small data update instead of regenerating every boundary. The populations sum to 775,610, which is the published provincial total exactly, so the join is complete and nothing was dropped.

The original GeoNB file arrives as fifteen features built from 636 separate rings, because every island in the Bay of Fundy and every sandbar off the Acadian Peninsula is its own record. Those dissolve into fifteen counties. Simplification then runs along shared arcs rather than county by county, which matters more than it sounds: simplify two neighbours separately and their common boundary drifts apart, leaving slivers of empty space between counties that should be touching. Done properly the shared boundary is processed once and both sides move together.

Why the boundary is not the same thing as the land

Ask two agencies how big Queens County is and you get 3,681 square kilometres from Statistics Canada and 3,946 from GeoNB. Neither is wrong. Statistics Canada reports land area. The GeoNB polygon is a boundary, and that boundary runs around Grand Lake and Washademoak Lake rather than excluding them, so it encloses 7.2 percent more ground than the county has land.

The same thing happens to Saint John at 5.4 percent, which encloses its harbour, and to York at 14.3 percent, which encloses the Mactaquac headpond. The tool shows both numbers side by side in the Selection panel with the difference spelled out, because someone checking a figure against another source should find the explanation on the page rather than assume the tool is broken.

Density is always computed from land area. Using polygon area would quietly understate every county with a big lake in it, and it would understate them unevenly, which is worse than being uniformly wrong.

Three counties run the other way, and only one of them matters. Kent and Restigouche come in 0.0 and 0.4 percent smaller, which is inside the noise of simplification. Northumberland is a different case. Its polygon measures 5.0 percent smaller than the reported land area, and enclosed water cannot produce that result in any quantity, because water inside a boundary can only push the polygon above the land figure, never below it. Probably a difference in how Miramichi Bay is assigned between the two agencies. It is flagged rather than smoothed over.

The projection, and why Web Mercator is the wrong answer here

Every web map you have ever scrolled is Web Mercator, and for Canada it is a poor choice. Mercator scale error grows with latitude, so Nunavut arrives roughly four times too large and a country that is mostly north gets systematically distorted.

This tool uses Lambert Conformal Conic, which is the standard projection for Canadian mapping, with the two standard parallels set by the one sixth rule: place them one sixth of the way in from the north and south limits of whatever you are drawing. For New Brunswick, spanning 44.50 to 48.07 degrees north once Grand Manan and the northern tip are included, that puts them at 45 degrees 06 minutes and 47 degrees 29 minutes, and the parallels recompute for each province rather than being fixed.

What it buys you is measurable. Straight line distances taken off the projected coordinates match great circle distances to better than one part in 4,600 across the province. Fredericton to Moncton comes out at 145.13 kilometres projected against 145.16 by great circle, an error of 0.02 percent, or about thirty metres over 145 kilometres. On a Mercator map at this latitude you would be out by a great deal more, and the error would grow as you moved north.

Satellite imagery, and how it is fitted to the map

The Satellite map type replaces the flat county colours with real imagery of New Brunswick. It comes from NASA, specifically the Blue Marble shaded relief and bathymetry layer served through Global Imagery Browse Services. That layer is a cloud free composite rather than a single day of observations, which is why you can see the whole province at once instead of the cloud bank that usually sits over the Bay of Fundy.

Fitting it to this map takes a step that most web maps skip. The imagery arrives in plate carree, where latitude and longitude are treated as a simple grid. This map is drawn in Lambert conformal conic, for the reasons in the section above. Laying one over the other directly would leave the coastline sliding away from the county boundaries as you move north, by several kilometres at the top of the province.

So the imagery is reprojected in your browser before it is drawn. Every pixel on the map is converted back to a latitude and longitude, and the matching pixel is taken from the source image. The result lines up with the boundaries because both are in the same projection. The imagery is then clipped to the provincial outline, so the map shows New Brunswick rather than a rectangle of the northeast.

The imagery is public domain. NASA asks to be credited rather than paid, and the credit is drawn onto the map itself so it travels with the image when you download it. This is also why the imagery is NASA rather than the sharper commercial layers you see on mapping sites: those are licensed from private operators and cannot be redistributed through a download button.

Two practical notes. The imagery takes a second or two to appear the first time, because it is fetched and reprojected on the spot. And at roughly 500 metres per pixel it is a landscape view rather than a street view, so you can make out the Saint John River valley, the forest cover and the coastline, but not individual buildings.

A check you can repeat on paper

Take Westmorland County, the Moncton end of the province and the most populous of the fifteen.

Statistics Canada gives it 163,576 people and 3,659.74 square kilometres of land. Density is one division:

163,576 divided by 3,659.74 equals 44.70 people per square kilometre.

Select Westmorland in the tool and the Selection panel reads 44.7 per square kilometre. Switch the units to imperial and it becomes 115.8 per square mile, which is the same number multiplied by 2.58999, the square miles in a square kilometre.

Now do it the wrong way, using the polygon area of 3,690 square kilometres instead. You get 44.33, which is 0.8 percent low. On Westmorland that is a rounding difference. Run the same mistake on Queens County and 2.99 per square kilometre becomes 2.79, an error of 6.7 percent, because Queens encloses two large lakes and Westmorland does not. That is why the denominator matters and why the tool never uses polygon area for density.

One more you can verify without any arithmetic: add the fifteen county populations from the table. They come to 775,610. If your total differs, a county was missed.

Mistakes that catch people out

Reading county area as land area. They differ by up to 14 percent here, and the gap is largest exactly where you would least expect a problem, in the counties with lakes and harbours. Always check which one a source is quoting.

Assuming county means the same thing across Canada. It does not. New Brunswick, Nova Scotia and Prince Edward Island have counties. Quebec has regional county municipalities, British Columbia has regional districts, Saskatchewan and Manitoba have rural municipalities, and Newfoundland and Labrador has no county layer at all. Census divisions are the one layer that exists everywhere, which is why this tool is built on them.

Treating the river courses as survey data. They come from the New Brunswick Hydrographic Network, filtered to stream order six and above and then generalised for display, so the courses are real but the detail is not survey grade. For full resolution hydrography go to the GeoNB watercourse layer directly.

Comparing densities across provinces using quantile shading. The bands here are fixed rather than quantile, so a colour means the same number in every province. Quantile shading recolours itself for each dataset, which makes two maps look comparable when they are not.

Expecting the county boundary to follow the shoreline exactly at high zoom. The geometry is simplified to 4,569 vertices for the whole province, which costs at most 0.17 percent of any single county area against the full resolution original. It also smooths harbours, which is why the city point for Saint John lands about a kilometre outside its own coastline and the tool falls back to the nearest county rather than reporting none. That is invisible at province scale and it is not a property survey.

Related Calculators and Reference

• Canada Map: the national view, with all thirteen provinces and territories and a link into each one that has its own map.
• Coordinate Converter, latitude and longitude to UTM: this map reads out in degrees, minutes and seconds, so go here when you need the same position as a UTM grid reference.
• GPS Waypoint Logger: collect points in the field, then bring the CSV back and check them against the county they fall in.
• Contour Map Generator from XYZ data: this page has no terrain layer yet, so use this when you have your own elevation points to contour.
• How to make a contour map from survey data: the background reading for turning raw levels into a surface before you map anything on top of it.
• Cross Section Maker from boreholes: once you have located your holes on a plan, this turns them into a section.
• Outcrop Photo Annotator: for adding a scale bar and north arrow to field photographs taken at the sites you have mapped.
• Statistics Canada 2021 Census boundary files: the official source for census division geometry across every province and territory.

County boundaries: Service New Brunswick, GeoNB, used under the Open Government Licence New Brunswick. Population and land area: Statistics Canada, 2021 Census of Population, used under the Statistics Canada Open Licence. Neither agency endorses this product.