Arc Collector Activity 1: Introduction

Introduction

This activity served as an introduction to Arc Collector, an extension of ArcGIS Online that allows users to collect and update data in the field, log current location, and map data in real time. With this application, maps can be created and used anywhere to ground-truth data, make observations, and respond to events. It is highly accurate and efficient and is a logical use of resources, since today’s smart phones and tablet devices have many times more computing power than most GPS units, often coupled with better internet access. Arc Collector’s practicality makes it an excellent app for burgeoning GIS professionals to be familiar with.

For more info on Arc Collector, visit the ArcGIS Online site here. 

In this activity, the class used Arc Collector to record microclimate data from around campus using smartphones and tablets. Our objectives were to collect dewpoint, temperature, wind speed and wind direction in different areas on campus and then use the data to map the microclimate of UWEC. We were to have at least one map for wind, dew point, and temperature.

Materials Used

•      Ipad
•      Kestrel 3000 portable weather station
•      Brunton compass (to collect wind direction)
•      iPhone 5s

Study Area

The study area was the main University of Wisconsin Eau Claire campus and Putnam Park, a state natural area incorporated into the campus. Campus is naturally divided into two sections by the infamous UWEC hill. Lower campus houses the academic buildings, library, and student center, and is divided by the Chippewa River. Upper campus contains the Hilltop Cafeteria, two athletic facilities, the student health center, campus police headquarters, and nine residence halls. Putnam Park includes the hill and the areas on either side of the river valley. In order to ensure somewhat even coverage of the campus, the class was split into small groups. Access was granted to an ArcGIS Online account with a shared map of the campus, subdivided into 5 zones, and each group was assigned one of the 5 zones to collect data within (Fig.1). Group 9 was Jackie and Amanda (myself) and we were one of the groups assigned to zone 5, which included the upper campus residence hall area, the hill, and a large expanse of lower Putnam Park by the Chippewa river.

Figure 1 The map of campus that was shared with the class via ArcGIS Online. Groups were assigned a zone to collect data within.

Methods

Group 9 set out with the Kestrel 3000, the compass, and Jackie’s phone (iPhone 5s). Amanda, lacking technology of her own, borrowed an iPad from the University Geography department. ArcGIS Online and Arc Collector were opened and the project was deployed. Each group was tasked with collecting 10 or more data points. Jackie and Amanda chose to take the Putnam Park trail and walk along the bottom of the hill and through the river valley area.

Data collection consisted of holding up the Kestrel 3000 until it could get a steady read on dewpoint, wind speed, and temperature (Fig.2). Wind direction was measured by facing the wind and then measuring the azimuth. These attributes were typed into the Arc Collector chart associated with each point and updated at each individual site. Each time an update was made by any of the groups walking around campus, the data point was added to the map so that it could be seen by the whole class. The iPad, in an unexpected twist, did not connect to the campus internet, and was rendered useless for data input. It was used instead for its camera during the excursion.

Figure 2 Jackie collecting the microclimate attribute data. On the left, the Kestrel 3000 collects dewpoint, wind speed, and temperature. The wind direction azimuth was calculated using the compass (right photo).

All ten points were collected with ease. In addition to Putnam Park on lower campus, we collected points on the hill, on the tennis courts, and near Towers Hall (Fig.3).

Figure 3 Action shots of Group 9 collecting data. On the left, Amanda reads dewpoint values in the Putnam Forest. Jackie poses with the Kestrel 3000 on a scenic portion of the Chippewa River and checks wind direction on upper campus.

Generating Maps

Now that all the data was compiled into ArcGIS Online, it was time to return to the lab and commence the mapping process.

ArcGIS Online (Fig.4) was accessed and opened in ArcMap desktop. I saved the data points to my newly made “MicroClimate” geodatabase as a point feature class “MicroPts_sengera”.

Figure 4 A screenshot of the ArcGIS Online dashboard.

I also saved the “Microzone” boundaries which designated the zones on campus that groups were assigned to collect data within. In order to make a boundary around the total study area, the Dissolve tool was used. This created a polygon with no divisions between zones. The editor tool was utilized to move the boundaries to include all of the points, since there were a few data points a little bit outside the boundary lines (Fig.5).

Figure 5 Results of the Dissolve and Editor tools.

The Feature to Raster tool was used to convert the temperature field in the “MicroPts_sengera” attribute table to a raster that would display temperature as a continuous surface. Unfortunately, it didn’t work. Instead of resulting in a smooth surface, it made discrete dots that were color-coded. Interpolation was necessary. Spline was the obvious interpolation choice for temperature, since it is best suited for continuous surfaces (see my Digital Elevation Surface blog post for a review of interpolation methods). The Spline Spatial Analyst tool was used on the temperature field. This resulted in a continuous surface map (Fig.6). The same Spline interpolation tool was used on the dewpoint data field.

Figure 6 A continuous surface map of the temperature data. This needed to be massaged for a better display, but the largest step was complete. 

The wind direction data proved more difficult to map. The spline interpolation tool was used on both wind speed and wind direction, then added to the map. Under Properties>Symbology on the wind direction feature, Vector Field was selected. This allows arrows to be used to indicate wind direction. Importing the wind speed data allowed for the arrows to be colored in accordance to a scale of magnitude (Fig.7).

Figure 7 Selecting Vector Field allowed the data to be displayed as Beaufort wind arrows. This demonstrates wind direction and magnitude simultaneously.

Finally, I made a more comprehensive map of the study area. This involved editing the microzone feature class; some vertices had to be moved around to eliminate gaps and overlaps of the zones.

Results and Discussion

Unfortunately, despite the division of the class in an attempt to capture information in a relatively even distribution across campus, no groups collected data in Zone 4, which left a large data gap in the center of campus. This data void was unfortunate and undoubtedly affected the overall accuracy of the microclimate assessments. Notice there is data void over the river and also in the left side of Zone 5 corresponding with the Towers Parking Lot. Data collection in the river was prevented by the lack of kayaks, but the lack of data points in the parking lot was an accidental oversight. Apart from these three areas, the data points are relatively uniformly distributed. 

Figure 8 The map of the UWEC campus with the five designated zones. The points show each data entry to the ArcGIS Online map. Notice the lack of data points in Zone 4. 

The next map compares temperature, dewpoint, and wind speed/direction data taken at each point. The data was interpolated into a continuous surface map for each attribute (Fig.9). 

Figure 9 This map compares the climate attributes that were collected. 

It is interesting to note that the river does not seem to affect the microclimate data. Since the data was recorded on a 60 degree (C) mid-November day, it would be expected that there would be at least a small low pressure system above the river. This would be expected to have obvious affects on dewpoint and temperature in the areas directly above the water, at least. Instead, temperature and dewpoint appear to vary just as much over the water as they do over the upper and lower campus areas. Just from looking at the data without the background map, the river would not be evident at all. This could be because there were no data points taken directly above the water and the interpolation tool averaged over that area. It is also interesting to note the high temperatures recorded above some buildings on lower and upper campus. Reported at 89 degrees, this could correspond to heat vents outside the buildings or even the ambient temperature inside the buildings, but it is unlikely that the temperature inside the buildings was that high, much less the area outside the buildings. Some sort of error must have occurred in the data there.

In order to best showcase the wind data, I decided to make a separate map for wind speed and direction, in addition to the one included in the main attribute map, so that I could include more detail (Fig.10).

Figure 10 This map demonstrates the wind direction and magnitude in meters/second reported across the campus. The arrows denote the direction the wind is reportedly coming from.

This map was interesting, since the amount and size of the Beaufort wind arrows appeared very differently in ArcMap before the map was exported. The arrows in the final map are small and difficult to see, which is an unfortunate situation I will work to remedy in the future, perhaps by collecting U and V data and creating a custom vector field. It is another point of contention with the Beaufort wind display that the legend would not allow me change the colors of the arrows or to truncate the legend. In the map of campus, the highest wind magnitude reported was an 14.1, but the pale yellow arrows blend in too well to be seen against the background map. I attempted to remedy the situation by making the background map transparent, but it didn't help much. 

Analyzing the wind data, there seems to be no distinctive pattern of wind speed or direction indicated on the map. This could be due to many factors; wind is blocked in many places on campus by the hill, Putnam Forest, and by buildings. It is also likely that students did not report wind direction in a standardized manner. Some may have been facing the wind, and some may have had their backs to the wind. This discrepancy may explain the chaotic air flows and vortexes reported on campus. 

Conclusions

Overall, this field activity was a successful introduction to data collection using portable weather stations, and Arc Collector and ArcGIS Online. Collecting points was a simple matter of walking to a location that seemed to be far away from any other group’s data points, taking the measurements, and then recording and updating the map. It was incredibly easy and efficient. It was very convenient to be able to see where other groups were inputting data points in real time, it allowed us to disperse more evenly and collect a wider range of data.