Eohjin Lee

Date of Award

January 2020

Document Type


Degree Name

Master of Science (MS)


Geography & Geographic Information Science

First Advisor

Bradley Rundquist


The purpose of the study is to map agricultural drainage systems (ADS) at the watershed scale using remote sensing and GIS techniques and examine the effect of ADS. For achieving the purpose of the study, this study selected the Red River Valley (RRV) of the North in which agriculture is a primary industry at the region. Excessive nutrients, sediment, and pesticide from this agricultural area flow into the Red River throughout subsurface drainages. The ADS aims to remove excessed water from agricultural fields, and this ADS is divided into two systems - uncontrolled drainage system (UCDS) and controlled drainage system (CDS). While UCDS allows water flows to the stream or river through using pipes without controlling water table, CDS regulates water table by an equipped structure that controls the volume of water flows in the agricultural fields.

For mapping artificially drained tiles between UCDS and CDS fields in the RRV, this study used DEM to digitize linear, map slope, and to calculate surface area. This study digitized linear maps with eight UCDS and twenty CDS fields and the map contains a digitization of tile drainage locations, drainage system patterns, ADS types, and artificially drained surface areas. In the analysis of the two different ADS systems – i.e., UCDS and CDS, this study obtained the indexes by using NDVI, NDWI, and MSAVI2 provided by PlanetScope imagery. In testing the group difference between UCDS and CDS in the three different indexes is examined by computing Analysis of Variance Analysis (ANOVA). Also, this study postulated CDS is more effective to the healthiness of crops than UCDS does in the ADS system.

The results of ANOVA indicated that there is no difference in the spectral indexes analyzed by ADS type (i.e., NDVI, NDWI, and MSAVI2). This result implies that the healthiness of crops is not affected ADS type, at least for the year studied here. The causes of this result, as a limitation, is derived from missing information, which is that agricultural research should consider region-specific crop calendars that involve significant idiosyncratic information, such as crop types and cycles, and regional climates. Nonetheless, natural factors in the RRV in 2019 – e.g., weather – are, of course, things outside of the control of researchers and agricultural producers.