Laserfiche WebLink
MEMO -TASK 90.2 <br />• The Landsat-derived NDVI composite classification was used to identify irrigated lands and the <br />DOQQ-derived image base maps were used to delineate boundaries of irrigated fields. Fields <br />with the same NDVI signature typically were mapped as one field but linear features on the <br />DOQQs determined the boundary. <br />• Irrigated Field boundaries were delineated using distinct physical linear features on the DOQQs, <br />including: roads, fence lines, canals and main ditches, row direction, changes in crop type, and <br />distinct changes in photo tone or texture. <br />• Field ditches/laterals that appeared to serve lands from both sides of the ditch were not used to <br />differentiate field boundaries but were mapped as a single parcel. <br />• Sharp changes in photo tone on DOQQs were not used to differentiate parcel boundaries unless <br />other information supported the separation of parcels. Within a field, there were sometimes <br />patches with dark DOQQ image color adjacent to patches with light image color. These <br />conditions were interpreted as harvesting practices at the time the imagery was taken or different <br />crop types in the year of photo acquisition, which was usually different from the 2001 reference <br />season. In these cases, if there was no physical structure dividing these areas then the Landsat- <br />NDVI classification was used for interpretation. If the irrigated field had the same, uniform <br />color on the Landsat-NDVI composite then it was mapped as one irrigated field. If the irrigated <br />field had a color change in the Landsat-NDVI composite at the same location as the tone change <br />on the DOQQ then it was mapped as separate irrigated parcels. <br />• Trails, driveways, and field access roads with width of 10 to 12 m or more were used to <br />differentiate parcel boundaries. If the crop appeared similar on the DOQQ and on the NDVI <br />composite, then the parcel was not separated. <br />• Hay fields with a similar NDVI composite signature were not separated if only divided by minor <br />ditches or other minor features. <br />• Irrigated homesteads were identified as rural housing developments with irrigated pasture or <br />hayfields mixed with houses and other structures. The individual homesteads were grouped and <br />mapped as a single parcel, and attributed in the database as "irrigated ranchette." <br />• The minimum mapping unit for an irrigated parcel was approximately 3 acres. This parcel size <br />accommodated most of the parcels in the study area. Fields smaller than this size were usually <br />lumped with an adjacent parcel or excluded if there was no adjacent parcel. Some exceptions to <br />this are corner fields of parcels irrigated by center pivots, which were sometimes less than 3 <br />acres. <br />• Minimum distance between field boundaries was established as approximately 12 m. This is the <br />approximate width of larger roadways and major ditches. <br />• Center pivots were typically mapped as true circles or semi-circles. Exceptions to this guideline <br />were made if the shape of a center pivot's coverage was disrupted by an obvious feature such as <br />a lake, non-irrigated soils, buildings or other non-irrigated features. Corners of center pivots <br />were typically mapped as separate parcels. <br />3. QAQC of Field Boundary Drawings. After field boundaries were manually drawn, they were <br />checked for accuracy and consistency with the guidelines described above. Parcels were edited <br />and revised as necessary. <br />4. Digitizing. Irrigated parcel boundaries were manually digitized on-screen using ArcMap. A <br />technician used the manually drawn irrigated parcel boundaries for digitizing, on-screen, using as <br />abase the detailed DOQQ imagery displayed on the computer screen at a scale of approximately <br />1:5,000. The technician attributed each irrigated parcel boundary with a polygon attribute <br />including: fallow, irrigated ranchette, center-pivot, or all other fields. In addition, each parcel <br />was attributed as either surface irrigated or sprinkler irrigated. Center pivots were considered to <br />be sprinkler-irrigated and other fields were identified as surface irrigated unless other field <br />information was available, such as field observation of side-roll sprinkler systems. <br />Page 4 of 6 ~l~lverslde Technology, lnc. <br />Gr:HS•: Rx:soarcc~s En;:ncer~nc.~~ad cuns.nrmc <br />