Out of the Fog
Humboldt Bay and the Eel River are located along the Pacific Coast, some 250 miles north of San Francisco, Calif. Beyond its natural beauty, this area provides valuable commercial, recreational and cultural resources for the northern part of the state. The region is an estuarine system containing diverse habitats including rivers; brackish and salt marshes; eelgrass meadows; tidal flats; large benthic areas of sand, mud and gravel; and sand and rock beaches. Although Humboldt Bay is the second largest estuary in California, it has received relatively little scientific attention to management, conservation and restoration of intertidal habitats. Prior to this project, the last habitat inventory and mapping was completed between 1970 and 1980. The Eel River Estuary is even less understood, and few studies of any kind have occurred there.
Conservation and management efforts linking land and sea have been limited by a relative lack of geospatial data for intertidal and subtidal habitats. Consequently, quantified habitat mapping from new high-resolution imagery was considered a fundamental need for an integrated ecosystem approach.
Several years ago, the NOAA Coastal Services Center (NOAA CSC) and the California Sea Grant Program came together to embark on an aggressive and challenging project to acquire new color digital imagery and to map the submerged habitat in the bay. Photo Science was selected for this project during the fall of 2006. Imagery was not complete, however, until the summer of 2009 because of the challenging weather conditions along the Pacific Coast combined with project requirements for acquiring the aerial imagery during very low tides.
The Humboldt Bay Initiative (HBI), an ecosystem-based management program, is an inclusive and diverse effort. The program is focused on a collaborative, sustainable scientific approach to natural resource management in the Humboldt Bay ecosystem and attempts to address issues such as endangered species conservation, degraded water quality, reduced salmonid populations and adapting to coastal climate change impacts. Program participants include over 100 professionals from more than 30 organizations, agencies and institutions, resulting in a community-based effort. This team has a strong network of community partnerships, high awareness and involvement in natural resource issues, a strong connection to the land, more than 40 years of restoration experience, and an understanding that the Humboldt Bay ecosystem is changing in unprecedented ways.
An inventory of existing mapping, benthic habitat and species databases was completed before acquisition of the new imagery was planned. Stakeholders identified priority data needs and assisted with completing the database at a public meeting and were individually contacted by Photo Science. The inventory included 116 GIS datasets and 12 non-GIS datasets with the type of data available, accessibility and the spatial extent described. Habitat-related data comprised 48 datasets, most of which were created after 2001. However, this inventory produced very little data that could be used to support the management efforts of the Humboldt Bay region. Accuracy and completeness were difficult to determine due to the relative lack of quality metadata describing the available datasets. Further, many datasets were developed well prior to the current management initiatives and it was unclear whether those data could support the current and projected conservation efforts for Humboldt Bay. Therefore, the stakeholders agreed that a new mapping effort was required to support both scientific and management efforts. This decision meant that new high-resolution aerial photography was required to properly map the benthic habitats of Humboldt Bay.
The water temperature of the Pacific Ocean in northern California is quite cool, ranging from about 49oF in winter to 55oF in late summer. The influence of the ocean moderates temperatures in the project area, providing for warmer winter and cooler summer weather. The area also experiences high relative humidity coupled with a prevailing wind from the northwest. The steady flow of moist, warm air in the summer from the northwest sweeps over the cool water along the coast, creating morning fog and low clouds along the coast. By late morning or early afternoon, the temperatures climb enough for the fog to burn off and the clouds dissipate.
For many other mapping projects, morning fog and low clouds are not problematic. Acquisition can simply wait until the early afternoon when the skies clear. However, this project included significant restrictions on tidal conditions for the imagery that were in direct conflict with the typical morning weather over the Bay. Further complicating the acquisition was the mixed tide characteristics of the California coast. With mixed tides, there are two high and two low tides each lunar day, but there is a large inequality between the heights of either the two low tides or two high tides. During the project, the lowest low tides always fell during the morning hours when the Bay was typically obscured by fog and low clouds. Many otherwise acceptable tide conditions occurred too early in the morning for imagery acquisition. Additionally, the project called for sun angles ranging from 25 to 40 degrees to minimize shadows in the imagery and glint, or the sun’s reflection off the water. For most of the acquisition window, the 25-degree minimum wasn’t met until sometime between 8:00 and 9:00 a.m. PDT.
The project was also challenging due to the rural nature of large portions of the project area. It is ideal to have numerous ground control points spread throughout the project area to properly control the aerial imagery. But access permission on some of the properties would have been very difficult to secure. Therefore, the decision was made to control the project using a combination of airborne GPS and inertial navigation systems during photo capture and to supplement this information with limited ground control points on photo-identifiable features confined to public rights-of-way. This control technique provided excellent accuracies in the digital orthophotography.
Finally, water clarity was also a big concern. The success of the photo interpretation was based both on low water elevations in the Bay during acquisition to expose some of the vegetation and on the ability to see into the water column to allow classification of the submerged aquatic vegetation. Environmental and atmospheric conditions within the project area were carefully monitored daily to ensure the acquisition took place well after any significant rainfall event in the area and under generally calm wind conditions. This last requirement ensured good water clarity.
After three years of monitoring weather conditions at the site for every possible flight date, and with the significant help of the zone forecasters at the NOAA National Weather Service in nearby Eureka, near-perfect weather conditions occurred. It took a rare combination of the right lunar conditions coupled with a high-pressure system sitting over the Bay that included an offshore wind component to provide the clear blue skies for a midmorning, low-tide acquisition. An aircraft and digital sensor were mobilized to the project site the night before and successfully completed the image acquisition on June 27, 2009.
Research had to be completed each year to determine the windows of opportunity for acquisition. These windows were communicated to HJW Geospatial, the Oakland-based flyer responsible for imagery acquisition so the firm could make flight preparations. (HJW Geospatial was later acquired by Photo Science.) Predicted tides at North Spit, a NOAA tide gauge located in Humboldt Bay, were downloaded from NOAA’s Center for Operational Oceanographic Products and Services (CO-OPS) Web site. Low tides of -1.0 foot or lower were added to a spreadsheet, along with the water level start and end times predicted to be below the threshold. These preliminary times then had to pass two additional tests. The first test was a cross check to determine if the tide window also met the requirements for the minimum sun angle. The second test determined whether this window was long enough to allow for the imagery acquisition. Normally the first and last days within a tidal opportunity are so limited that no real acquisition can take place. Table 1 lists a window of opportunity for April 2009, when one attempt at imagery acquisition took place.
Note the significant disparity between the heights of the morning and evening low tides in these mixed tides. Based on the table, it is evident that no times were available on the 25th and 26th where the tide was below -1.0 foot and the sun was at least 25 degrees above the horizon. Also, based on both tides and sun on the 27th, the window extended only from 8:38 to 9:27 a.m.--obviously not long enough to warrant an attempt, even if the weather had cooperated that day. So the only potential acquisition window for late April was the 28th and 29th. Of course, the most significant test was the actual weather conditions at the Bay during the time of the predicted acquisition window. Unfortunately, the site experienced typical fog and clouds on these two days in April, so the imagery was not captured. This played out time and time again over three years with planning taking place for every potential flight date until weather conditions finally cooperated.
Flight planning became an incredibly tedious task to carry out manually, with many opportunities for error in reading/recording tide height interpretations, time conversions between Greenwich Mean Time and local time zones, and tracking the effect of daylight savings time. With an understanding of these challenges along with recognizing the benefits that would accrue from an automated approach, the Photo Science team undertook the development of an automated tide tool for flight planning. They knew the value of such a tool would extend well beyond this project to a host of other tide-restricted imagery and LiDAR acquisition projects.
This tide tool reduced the tide planning effort for this project from about two hours per potential flight day to less than five minutes. It also greatly reduced the potential for mistakes made in tide planning, provided a digital audit trail of all decisions made during the flight planning, and allowed for easy verification of the acceptability of flight times versus actual tidal conditions. Since its development, Photo Science has successfully used it on other imagery and LiDAR projects undertaken on the Atlantic coast and Gulf of Mexico for the USGS, NOAA NGS, NOAA CSC, state agencies and regional commissions. Further, this tool will be used for mobile mapping applications in coastal communities in the very near future.
Habitat features mapped included unconsolidated sediments, coastal marsh, oyster mariculture, eelgrass, patchy eelgrass, macroalgae and subtidal. All habitat delineations were made with the highest precision possible to best reflect actual habitat boundaries on the ground. The minimum mapping unit for this mapping effort was 0.01 hectares (nominally an area 10 x 10 meters in size). Throughout the photointerpretation process, Photo Science collaborated with the NOAA CSC and California Sea Grant via Web conferencing to discuss and agree upon various photointerpretation calls. This allowed virtual meetings to take place with participation from knowledgeable professionals in California, South Carolina and Florida, with no travel requirements. Further, the project team met at Humboldt Bay to collectively acquire ground truth field data. The final map products met and exceeded the thematic accuracy requirement of 80 percent for each individual habitat type and 85 percent overall for baseline benthic habitat. Accuracy was determined by the development of an error (similarity) matrix, comparing field observations to photointerpreted habitat features (polygons).
The uses of the images and classification are diverse. California Department of Fish and Game uses them for natural resource management and restoration project design. For example, the imagery is used to infer vegetation types and the location of channels and fence lines. The USFWS uses include restoration planning; updating National Wetland Inventory mapping; invasive cordgrass and rare salt marsh plant mapping, management and control; and determining boundaries and acreages of habitat types. They also have plans to use the information to map changes in wetland types as a result of restoration projects. The USGS participates in a large federal program using the images for inter-federal agency planning and coordination and has downloaded them onto their Web site. The USGS system is connected to NASA and provides another mechanism for people to access the imagery through robust map services.
Humboldt State University (HSU) students use the images for various projects and thesis research. HSU faculty, and local consulting firms have used the imagery and eelgrass classification to identify changes in the Humboldt Bay eelgrass footprint. The California Coastal Conservancy and University of California Sea Grant Programs use the images to map invasive species such as Spartina densiflora and Zostera japonica. Sea Grant also uses the images in annual reports to funding partners and for permit applications. One local business is using the images to map Humboldt Bay shoreline infrastructure and quantify shoreline types. The Humboldt Bay Harbor, Recreation and Conservation District used the habitat classification when working with regional stakeholders on the North Coast Marine Life Protection Act process.
Several users say they expect to use the images for sea level rise projects, especially for climate adaptations planning in conjunction with coastal LiDAR data expected by the end of 2010. The Pacific Coast Joint Venture plans to use the images for conservation planning to help determine habitat goals.
High-resolution images contribute to improved scientific understanding of coastal habitats, promote effective management efforts, establish baselines for monitoring environmental change, give insight to shoreline processes and impacts of storms, supports tsunami, earthquake, and flooding hazard assessments. They are also excellent visualization and educational tools. This imagery and classification are publicly available and can be used as a baseline for impact and change assessments. Digital Coast is an information resource with a user-friendly interface and relatively jargon-free language to facilitate science based coastal decision making. It ensures that the information is available and connects scientific information to managers.
Photo Science was honored with the 2010 MAPPS Geospatial Products and Services Excellence Award in the Airborne and Satellite Data Acquisition category for its work in the Humboldt Bay Ecosystem Mapping Project. For more information about MAPPS, visit www.mapps.org .
NOAA CSC and the Coastal Community
The National Oceanic and Atmospheric Administration’s Coastal Services Center (NOAA CSC) was an integral partner in this effort. The center provides tools and services to those organizations who manage the nation’s coastal resources, including local governments, nonprofits, state organizations and other federal agencies.
One of the center’s goals is to provide new spatial data and information for the nation and the aforementioned constituents, which made the Humboldt Bay project a good fit. Projects like Humboldt Bay, which have local, active partnerships participating in the effort, are usually the most successful the center undertakes; having end users play an important role in gathering and producing data ensures data relevancy in local decision-making. At the same time, one role played by the center is to ensure that final data products can be easily shared with a broader, national audience.
To reach the national audience, the center developed an enabling platform to assemble coastal geospatial data, tools, training and information called The Digital Coast. The Digital Coast includes land cover, topography and bathymetry, aerial imagery, benthic habitat, point samples, and other relevant coastal data. Each of these data types is important to a holistic understanding of the coastal environment. The Web site goes a step further and provides access to the training, technical guidance, and management tools often needed to turn the data into information that is useful and pertinent to decision-making.